Page 45 - Read Online

Klaassens et al. Microbiome Res Rep 2024;3:38 https://dx.doi.org/10.20517/mrr.2024.13 Page 11 of 19

(data not shown). Upon colonic fermentation for 48 h, small differences in the overall composition of the
donor microbiota composition were observed for infants born via VD or CS [Figure 4A-C]. Family-,
Genus-, Species-, and Strain-level relative read abundance profiles were generated using shotgun
metagenomics sequencing. Though all infants were of approximately the same age and all were exclusively
breastfed, high inter-individual variability was observed in microbial diversity, as expected for this age
group [Figure 4], such as the difference in abundances of Bifidobacteriaceae and Clostridiaceae [Figure 4A].
Small effects of the formula matrix can be seen in some infants, but not in others.

The presence of Bb significantly reduced the mean relative abundance of the genus Salmonella within the
family Enterobacteriaceae (P = 0.0045, P = 0.0057), the genus Streptococcus within the family
Streptococcaceae (P = 0.0091, P = 0.0098) and the family Lachnospiraceae (P = 0.0044). The reduction of
[40]
these genera and families was associated with a reduction in virulence genes in the metagenome .
Probiotic abundance in the in vitro colon model after 48 h of fermentation
The analysis pipeline with a specific infant database, as created here, enables taxonomic classification at the
strain level in infant microbiota samples. This approach was applied to track the probiotic strains in the
microbiota samples from the ex vivo model. The abundances of the probiotic strains supplemented in this
study are shown in Figures 5A and B. The observed abundance of Bifidobacteriaceae (relative abundance
average 33.80%) in the infant donor microbiota was much higher than Lactobacillaceae (relative abundance
average 4.47%) despite the same amount of supplementation. Each of the probiotics was detected in the
samples where it was supplemented, as expected.

In addition to the supplemented Bb strain, the infant microbiota also contained commensal bifidobacterial
species [Figure 4C]. In donor VD2 and donor CS2, a non-significant shift after the addition of the probiotic
strain was observed (light blue). The addition of Bb replaces commensal strains (e.g., B. catenulatum in
CS2) while promoting the growth of others (e.g., B. bifidum BGN4 in VD2). Interestingly, in all infants
except for CS2, Bb abundance was higher in combination with the eHF-GOS formula, despite its lower
growth rates in the in vitro setup. This suggests possible formula matrix-dependent effects as well as cross-
feeding mechanisms with other members of the infant gut microbiota. This is observed when Bb is added
both alone and in combination with Lf.

Addition of Bb indicates effects on the infant microbial community ex vivo
Fermentation for 48 h showed a trend toward an increase in bifidobacteria in all infants, independent of the
formula matrix [Figure 4C and Figure 5A], especially when supplemented with probiotic Bb; however,
changes were not significant (P > 0.05) [Figure 4A]. Interestingly, strain-level detection of Bb did show a
significant increase due to probiotic supplementation vs. no Bb supplementation. Bb-strain abundance
(27.31 median relative abundance) showed a significant (P-value < 0.005) shift from the samples without Bb
(0.46 median relative abundance), which is not observed with Lf supplementation. Interestingly, both
probiotics were added in the same amount. Principle Coordinate Analysis (PCA) was performed [Figure 6]
to visualize the impact of both probiotics on the overall microbiota. PERMANOVA analysis of samples
containing Bb showed a significant (P-value = 0.005) shift from the samples without Bb, which was not
observed with Lf supplementation alone, thereby consolidating the previous observation.

DISCUSSION
Our results show that in vitro, hydrolyzed infant formula (eHF) appears to have a strong influence on the
metabolic activity of probiotic strains. Interestingly, in the presence of a complex microbial ecosystem in an
ex vivo infant gut model, no effect on the metabolic activity of the microbiota was observed. However,