L. fermentum CECT5716 treatment ameliorated the gut
dysbiosis
The effects of L. fermentum on the intestinal microbiota were
explored since gut dysbiosis plays a key role in metabolic disorders,
including obesity. Several ecological features of the gut bacterial
communities were evaluated in the three experimental groups by different
parameters including Chao1 richness (diversity estimation), Phylogenetic
diversity (PD) whole tree (consider the phylogeny to estimate diversity
across a tree), Observed OTUs (count of unique OTUs in each sample) and
Shannon diversity (a richness and evenness estimator). Microbial
richness, evenness and diversity were significantly decreased in the HFD
group compared to the control diet group, whereas L. fermentumadministration was able to restore all these ecological parameters to
normal values, excepting Shannon index. Although it reached normal
values, it did not differ from the control group (Figure 7A).
Furthermore, the principal coordinates analysis (PCoA) showed evident
differences between control diet- and HFD-fed groups, thus indicating
two extremely different gut environments (Figure 7B). Of note, whenL. fermentum was administered to obese mice, marked differences
could be appreciated in comparison with non-treated obese mice (Figure
7B). To further investigate this remarkable shift in the gut microbial
environment after probiotic treatment, the bacteria composition at
phylum level was examined (Figure 7C). The 16S rDNA analysis revealed
that the most abundant phyla were Bacteriodetes (B) andFirmicutes (F) in all experimental groups, although HFD induced a
dramatic shift in both phyla when compared to non-obese mice,
significantly increasing Firmicutes (from 57.66% to 78.68%) and
reducing Bacteroidetes (from 23.72% to 13.02%). This resulted
in an increased F/B ratio, which was significantly restored by L.
fermentum treatment, reaching similar values to control diet-fed mice
(Figure 7). Similarly, Verrumicrobia phylum also showed a reduced
proportion in HFD-fed mice, which was significantly increased withL. fermentum administration (Figure 7C). Interestingly, the
relative abundance of Erysipelotrichi (class) was enriched in
untreated obese mice in comparison with non-obese mice, being this
situation ameliorated after probiotic treatment (Figure 7C).
At genus level, control obese mice revealed a reduced proportion in the
sequences in two genera, Bacteroides and Akkermansia ,
belonging to Bacteriodetes and Verrumicrobia ,
respectively, and an increase in Clostridium , in comparison with
control mice, which was significantly ameliorated after L.
fermentum treatment (Figure 7C).
Correlation of major bacteria genera and main obesity-related markers
using the non-parametric test of Spearman suggested that beneficial
effects observed in L. fermentum -treated mice could be derived
from microbiota modulation. In fact, a positive correlation betweenAkkermansia genus abundance and adipokines and intestinal barrier
markers expression, as well as, genus derived fromErysipelotrichaceae and inflammatory markers expression were
found (Figure S1). Thus, increased abundance of Akkermansia and
reduced presence of Erysipelotrichi (class) observed afterL. fermentum treatment could be involved in its beneficial
effects.
The PICRUSt (Phylogenetic Investigation of Communities by Reconstruction
of Unobserved States) analysis of the KEGG orthologs associated with
these bacterial signatures also showed different clusters among the
three groups (Figure 8A). LEfSe bar graph indicated that non-obese mice
group was characterized by a significantly higher (p<0.01)
proportion of bacterial taxa associated with the energy metabolism,
transport and catabolism, metabolism of cofactors and vitamins,
replication and repair processes and other glycan degradation (Figure
8A). In contrast, HFD mice harbored a higher (p < 0.01)
proportion of bacterial taxa associated with the cellular antigens,
biosynthesis of lipopolysaccharides and fatty acids, membrane
transporters and metabolism of amino sugar and nucleotide sugar (Figure
8A). When we evaluated the L. fermentum treatment in obese mice
bacteria related with glycan biosynthesis and metabolism, arginine and
proline metabolism, membrane and intracellular structural molecules,
nucleotide metabolism or secondary metabolism biosynthesis were
overrepresented (p < 0.01), which means that all of these
pathways were less represented (p < 0.01) in the other two
experimental groups of mice (Figure 8A).
Finally, we estimated the relationship between taxonomic and functional
enrichments in each group using the non-parametric test of Spearman’s
rank correlation. Results, as depicted in the form of hierarchical heat
map, showed distinct clusters of positive and negative associations
between bacterial taxa and metagenomic functions (Figure 8B). To further
simplify these clusters, we extracted the significant correlation
subsets (Spearman’s rho > 0.7 and p-value < 0.01)
and built separate correlation networks for each experimental group
(Figure 8C). Moreover, different co-occurrence networks arrays were
observed between non-treated and treated obese mice, confirming previous
results.