- Top of page
- Materials and methods
Aims: To evaluate the diversity of the Lactobacillus group in breast milk and the vagina of healthy women and understand their potential role in the infant gut colonization using the 16S rRNA gene approaches.
Methods and Results: Samples of breast milk, vaginal swabs and infant faeces were aseptically collected from five mothers whose neonates were born by vaginal delivery and another five that had their babies by caesarean section. After polymerase chain reaction (PCR) amplification using Lactobacillus group-specific primers, amplicons were analysed by denaturing gradient gel electrophoresis (DGGE). Clone libraries were constructed to describe the Lactobacillus group diversity. DGGE fingerprints were not related to the delivery method. None of the species detected in vaginal samples were found in breast milk-derived libraries and only few were detected in infant faeces.
Conclusions: The bacterial composition of breast milk and infant faeces is not related to the delivery method.
Significance and Impact of the Study: It has been suggested that neonates acquire lactobacilli by oral contamination with vaginal strains during delivery; subsequently, newborns would transmit such bacteria to the breast during breastfeeding. However, our findings confirm, at the molecular level that in contrast to the maternal vagina, breast milk seems to constitute a good source of lactobacilli to the infant gut.
- Top of page
- Materials and methods
Breast milk provides all the nutritional requirements for the rapidly growing infant and contains a variety of protective factors, such as immunoglobulin (Ig) A, immunocompetent cells, fatty acids, oligosaccharides, lysozyme or lactoferrin (Newburg 2005), that protect breast-fed infants against infectious diseases (Wright et al. 1998; Hanson and Korotkova 2002; Morrow and Rangel 2004). Commensal bacteria usually present in breast milk of healthy mothers, such as lactobacilli, lactococci, enterococci and Leuconostoc spp., can also be considered as key elements of the defence system that this biological fluid offers to the infant (Matsumiya et al. 2002; Heikkilä and Saris 2003; Martín et al. 2003). The lactic acid bacteria isolated from human milk include Lactobacillus gasseri, Lactobacillus rhamnosus, Lactobacillus plantarum, Lactobacillus fermentum or Enterococcus faecium, which all are considered to be among the potential probiotic bacteria. In fact, some of the lactic acid bacteria strains with such origin have already been shown to possess probiotic properties, including the inhibition of a wide spectrum of infant pathogenic bacteria by competitive exclusion and/or through the production of antimicrobial compounds, such as bacteriocins, organic acids or hydrogen peroxide (Heikkilä and Saris 2003; Beasley and Saris 2004; Martín et al. 2005).
Till now, knowledge of the bacterial diversity in breast milk is very limited, and is almost exclusively based on the use of culture media. This implies that the presence of additional bacterial species that are not cultivable may have been overlooked. However, the application of culture-independent molecular techniques, particularly those based on 16S rRNA genes, has allowed a more complete assessment of the biodiversity of the human microbiota, especially that of the gut both in adults (Zoetendal et al. 1998; Suau et al. 1999) and infants (Favier et al. 2002). Typically, this approach involves extraction of the DNA from the biological samples, polymerase chain reaction (PCR) amplification of 16S rRNA gene fragments with universal or group-specific bacterial primers, analysis of PCR products by fingerprinting methods such as denaturing gradient gel electrophoresis (DGGE) and, the construction of clone libraries to assess the variety of 16S rRNA gene sequences present.
As lactobacilli and other lactic acid bacteria seem to be important components of the breast milk microbiota, the primary objective of this work was to investigate their diversity in breast milk of healthy women by using 16S rRNA gene primers targeting the Lactobacillus group, which additionally comprises the genera Leuconostoc, Pediococcus and Weissella. Furthermore, the potential role of vagina and breast milk as a source of lactobacilli for the initial colonization of the infant gut was also investigated.
- Top of page
- Materials and methods
DNA isolation and subsequent PCR–DGGE analysis of 16S rRNA genes were successful for all samples. Visual comparison of DGGE profiles obtained from breast milk samples revealed interindividual variation, and no distinctive differences could be established between samples from the two different delivery groups. Similar results were observed for the samples obtained from vaginal swabs and infant faeces (Fig. 1). Cluster analysis of DGGE profiles by the UPGMA method based on the Dice similarity coefficient confirmed that no delivery-specific grouping could be retrieved (Fig. 2).
Figure 1. Denaturing gradient gel electrophoresis (DGGE) analysis of the dominant bacterial communities in breast milk and vaginal samples from mothers 3, 4, 9 and 10 and in infant faeces from their respective infants. M represents a marker constructed in this study with the identified bands to facilitate the interpretation of the figure. Bands: 1, Weissella confusa; 2, Leuconostoc citreum; 3, Lactobacillus gasseri; 4, Lactobacillus jensenii; 5, Lactobacillus iners; 6, Leuconostoc fallax; 7, Lactobacillus crispatus; 8, Lactobacillus fermentum; 9, Lactobacillus rhamnosus; 10, Lactobacillus casei and Lactobacillus paracasei; 11, Lactobacillus plantarum (fuzzy band).
Download figure to PowerPoint
Figure 2. UPGMA dendrogram illustrating the correlation between the different denaturing gradient gel electrophoresis (DGGE) profiles obtained from the samples (B, breast milk; V, vaginal swab; F, infant faeces) provided by the participating mother–infant pairs. The black bars represent the error bands.
Download figure to PowerPoint
Analysis of clone libraries prepared from PCR products obtained from the samples provided by mother–infant pairs 3 and 4 (caesarean section group) and 9 and 10 (vaginal delivery group), was used for further evaluation of the Lactobacillus group diversity. Sequence analysis of unique clones of these origins resulted in sequences with similarities >98% to 16S rRNA genes of cultured bacterial isolates deposited in the NCBI database. A total of 3, 4, 3 and 3 types of operational taxonomic units (OTU) were retrieved from breast milk of mothers 3, 4, 9 and 10, respectively (Table 1). On the other hand, 3, 2, 2 and 4 types of sequences could be retrieved from vaginal swabs of the respective mothers and 5, 4, 2 and 4 types of sequences from faeces of their respective infants (Table 1, Fig. 1). The Lactobacillus group patterns seemed to be host-specific and, globally, sequences belonging to 14 Lactobacillus group species could be identified (Table 1).
Table 1. Specific cloned sequences of the Lactobacillus group from breast milk, vaginal swab and infant faeces and the percentage of similarity to sequences deposited in the Genbank
|Sample||Individual*||Species†||Clone (% similarity)|
|Breast milk||M3||Lactobacillus fermentum strain BFE 6618||1BM3 (98)|
|Weissella confusa strain BJ21-2||2BM3 (98)|
|Leuconostoc citreum IH22||3BM3 (99)|
|M4||Lactobacillus rhamnosus strain MCRF-412||1BM4 (99)|
|Lactobacillus plantarum strain BJ G26-4||2BM4 (99)|
|W. confusa strain BJ21-2||3BM4 (99)|
|Leuc. citreum IH22||4BM4 (99)|
|M9||Lact. fermentum strain BFE||1BM9 (99)|
|W. confusa strain BJ21-2||2BM9 (99)|
|Leuc. citreum IH22||3BM9 (99)|
|M10||W. confusa strain BJ21-2||1BM10 (100)|
|Leuconostoc fallax strain BFE||2BM10 (99)|
|Leuc. citreum IH22||3BM10 (99)|
|Vaginal swab ||M3||Lactobacillus jensenii strain BJ H41-2b||1EV3 (99)|
|W. confusa strain BJ21-2||3EV3 (100)|
|Lactobacillus iners clone FX43-4||4EV3 (99); 5EV3 (97)|
|M4||Lactobacillus jensenii strain BJ H41-2b||1EV4 (99)|
|Lactobacillus crispatus strain BJ Y20||2EV4 (99)|
|M9||Lactobacillus iners clone FX43-4||1EV9 (99); 4EV4 (97)|
|Lact. crispatus strain BJ Y20||2EV9 (99)|
|M10||Lact. jensenii strain BJ H41-2b||1EV10 (99)|
|Lact. iners clone FX43-4||2EV10 (99); 6EV10 (97)|
|Lact. crispatus strain BJ Y20||3EV10 (99)|
|Aerococcus sp.||9EV10 (99)|
|Infant faeces||F3||Lactobacillus casei strain BJ13-1||1 F3 (99)|
|Lactobacillus paracasei isolate 10C||2 F3 (99)|
|Lact. paracasei strain VUP 12006||3 F3 (98)|
|Leuc. citreum IH22||4 F3 (99)|
|Lactobacillus gasseri strain YDB21||5 F3 (98)|
|W. confusa strain BJ21-2||6 F3 (99)|
|F4||Lact. plantarum strain BJ16-28||1 F4 (99)|
|Leuc. citreum IH22||2 F4 (99)|
|Lact. crispatus strain BJ Y20||3 F4 (99)|
|Lact. fermentum strain BFE||4 F4 (99)|
|F9||W. confusa strain BJ21-2||1 F9 (99)|
|Lact. fermentum strain BFE 6618||3 F9 (99)|
|F10||W. confusa strain BJ21-2||1 F10 (100)|
|Lact. crispatus strain BJ Y20||3 F10 (100)|
|Lact. gasseri strain KC5a||4 F10 (99)|
|Leuc. citreum IH22||5 F10 (99)|
Among the vaginal samples, most of the retrieved sequences were most closely related to 16S rRNA genes of Lactobacillus jensenii, Lactobacillus iners and Lactobacillus crispatus, but none of these species could be detected in breast milk-derived libraries, and only between 0 and 1 in clone libraries from the faeces of each infant (0, 1, 0 and 1 species in infants 3, 4, 9 and 10, respectively) (Table 2). For example, only one of the three Lactobacillus group species detected in the library corresponding to the vaginal swab of mother 3 was also detected in the infant faeces of her son. The ratio was lower (1 : 4) in the case of the mother–infant pair number 10 (vaginal delivery group). In contrast, two of the three species detected in the breast milk samples of each mother were also present in the faeces of their own infant (Table 2). Therefore, it is not strange that vaginal profiles of mothers 3, 4, 9 and 10 clustered together in the UPGMA dendrogram separately from the breast milk and the faeces profiles (Fig. 1).
Table 2. Comparison of the Lactobacillus group clones retrieved from breast milk (B), vaginal swab (V) and infant faeces (F) of the different individuals (3 and 4, caesarean group; 9 and 10, vaginal delivery group) included in this study
|Lactobacillus casei ||−||−||−||−||−||−||−||−||+||−||−||−|
- Top of page
- Materials and methods
Traditionally, the description of the bacterial microbiota of human mucosal ecosystems has been based on culturing techniques which tend to underestimate the bacterial diversity of such ecosystems (Vaughan et al. 2002). The use of molecular methods that rely on culture-independent approaches allows a more complete and reliable assessment of bacterial diversity (Vaughan et al. 2005). The 16S rRNA gene approach used in this study, DGGE analysis of PCR products obtained with Lactobacillus group-specific primers, results in a fingerprint that represents the diversity of the rRNA gene nucleotide sequences related to the species in this group (Favier et al. 2002; Heilig et al. 2002). The construction of clone libraries further allows the phylogenetic affiliation of populations corresponding to the bands which are visible within the DGGE patterns.
The PCR–DGGE profiles of all samples did not reveal a delivery-specific clustering. In fact, UPGMA analysis based on the Dice similarity coefficient revealed that the sample profiles from the different mothers clustered independently from the delivery mode. The origin of the lactic acid bacteria present in breast milk and in the infant gut is far from clear. It has been suggested that neonates would acquire them by oral contamination with vaginal strains during the transit through the labour channel; subsequently, newborns would transmit such bacteria to the breast during breastfeeding (Mackie et al. 1999; Isolauri et al. 2001; Heikkilä and Saris 2003). The Lactobacillus phylotypes detected in the vaginal samples are in agreement with recent 16S rRNA gene-targeted culture-independent studies, in which Lact. crispatus, Lact. jensenii and Lact. iners were identified among the dominant species in the vagina of healthy women (Burton et al. 2003; Zhou et al. 2004). Interestingly, none of the Lactobacillus species detected in vaginal samples were present in breast milk provided by the women whose neonates were born by vaginal delivery, a fact that suggests that transit through the vagina does not play a role in the establishment of lactobacilli found in breast milk. Additionally, the profiles of Lactobacillus sequences retrieved from infant faeces were more similar to those retrieved from breast milk of the respective mothers than to those obtained from the respective vaginal swabs (Table 2, Figs 1, 2). Such results are also in agreement with recent molecular studies which have shown that bacterial colonization is not significantly related to the delivery method and that vaginal delivery has, if any, a minor role in the development of the infant gut microbiota (Tannock et al. 1990; Matsumiya et al. 2002; Martín et al. 2003; Ahrnéet al. 2005). A molecular epidemiological study on the transmission of vaginal Lactobacillus species from mother to the newborn infant showed, that only less than one-fourth of the infants acquired maternal vaginal lactobacilli at birth, and that 1 month later, they had been replaced by lactobacilli associated with human milk (Matsumiya et al. 2002). It is illustrative that although the infant faeces used in this study were obtained at day 7 after birth, the spectrum of Lactobacillus group species found in such samples closely resembled that of the respective breast milk samples and not the respective vaginal swabs (Figs 1, 2).
It is not strange that the spectrum of Lactobacillus group sequences retrieved from breast milk and infant faeces was narrow, as within a mother–infant pair, the Lactobacillus composition of infant faeces and breast milk seems to be host-specific and usually only includes a low number of lactobacilli strains (Heikkilä and Saris 2003; Martín et al. 2004). As an example, the examination of Lactobacillus colonization in 112 infants during the first 6 months of life showed that 26% of them lack lactobacilli, 37% carried a single strain, 26% two strains and only 11% three strains or more (Ahrnéet al. 2005). Sequences matching with Lact. gasseri were not found in this study, whereas culturing-based studies have described it as a frequently found breast milk species (Matsumiya et al. 2002; Martín et al. 2003). This may be explained by the low number of mothers (n = 4) whose samples were used to construct clone libraries and/or possibly to its presence as a subdominant species within the Lactobacillus species spectrum existing in the participating mothers. At present, work is in progress to elucidate, at the strain level, if there is a mother-to-child vertical transmission of lactobacilli.