Page 28 - Read Online

Page 10 of 15 Ladeira et al. Microbiome Res Rep 2023;2:9 https://dx.doi.org/10.20517/mrr.2023.01

[35]
[34]
identify the subjects most likely to respond to dietary interventions or medical treatments .
Bifidobacterium is a common member of the human gut microbiome, with different species co-existing in
the host at different ages . A number of studies have shown that various metabolic, immune, and intestinal
[36]
[1]
disease states coincide with the depletion of Bifidobacterium from the gut microbiota . Here, we studied the
variability of the gut microbiome as a function of the resident Bifidobacterium community, using a public
database compiling curated metagenomics-based studies, mostly performed in adults from Western
countries. We first checked that we could reproduce the previously reported findings of a lower abundance
[1]
of Bifidobacterium species in antibiotic users [31-33] and individuals with diseases or adopting a non-
westernized lifestyle , the differential prevalence of most Bifidobacterium species depending on age , the
[28]
[10]
[27]
high prevalence of B. longum throughout the human lifespan , and the higher prevalence of B. adolescentis
in healthy subjects than in those with diseases [29,30] .
We then studied the variability of Bifidobacterium community composition with the Dirichlet multinomial
mixtures (DMM) method, which has been used for gut microbiome clustering on the basis of composition
in many studies [24,37-42] . We identified partitions enriched in different combinations of Bifidobacterium
species in a database containing predominantly adult data. The partitions that were more prevalent were
characterized by a higher abundance of B. longum and B. adolescentis, whereas the partition corresponding
to the non-detection of Bifidobacterium was the least prevalent. A specific analysis of three individual
cohorts of adults (~1,000 subjects) confirmed the detection of several partitions, indicating an effect of
between-subject variability rather than technical differences between studies. Those associated with a
healthier state were dominated by B. longum and B. adolescentis. In previous studies, the species-level
analysis revealed a positive correlation or covariation between multiple Bifidobacterium species [43,44] or
between specific species, such as B. adolescentis and B. longum [45-47] or B. adolescentis and B. bifidum .
[45]
However, another study reported a negative correlation between B. adolescentis and B. longum ,
[48]
suggesting variability between studies or study subjects. It remains unclear whether positive correlations
indicate metabolic cross-feeding or similar niches, and this aspect requires further investigation in vitro.
Cross-feeding between Bifidobacterium species on human milk oligosaccharides has been studied for the
species prevalent in infants [49-51] , but, to our knowledge, there have been no studies investigating cross-
feeding on complex dietary fibers between B. longum, B. adolescentis and B. pseudocatenulatum, which are
more common in adults.

The partitions with a higher abundance of Bifidobacterium, and particularly those dominated by both B.
longum and B. adolescentis, were associated with a higher gut microbiome diversity and a higher abundance
of butyrate-producing species, including Roseburia faecis, Coprococcus catus, C. eutactus, C. comes, and
Eubacterium hallii. Covariation between Bifidobacterium species and other resident species has been
detected for butyrate producers in the metagenomic analysis [43,46] , and metabolic interactions between B.
adolescentis, B. longum, and butyrate producers have been observed in vitro in the presence of complex
dietary substrates [52-54] . These partitions were more frequently found in healthy subjects, suggesting that
individual stratification exclusively on the basis of gut Bifidobacterium species abundance is associated with
differential gut microbiome structure and state of health. Notably, the partition corresponding to an absence
of Bifidobacterium detection, which contained a larger number of subjects with a non-westernized lifestyle
than the other partitions, was depleted of Streptococcus salivarius, which is detected in consumers of
yogurts, including yogurts supplemented with B. animalis subsp. lactis . Overall, our results extend
[55]
previous findings on associations with the gut microbiome by differentiating gut microbiomes enriched in
specific Bifidobacterium types. It would be interesting to determine whether these partitions are associated
with differential gut microbiome permissivity to distinct exogenous Bifidobacterium species/strains (during
and/or after the cessation of consumption) [56-58] .