The main aim of the present study was to determine whether antimicrobial food additives administered at the exposure levels reached in EU populations can trigger the dysbiosis of human gut microbiota and also whether the host genotype specifically Nod2-deficiency has any effect on the gut microbiota’s susceptibility to additives.
To accomplish this goal, we have colonised germ-free C57BL/6 mice with gut microbiota isolated from a healthy human donor. These human microbiota-associated (HMA) experimental mice of wild-type and Nod2-/- genotype, an animal model for Crohn’s disease, were provided with either sterile water or water supplemented with a mixture of the most commonly used antimicrobial food additives (AMFA) including sodium benzoate, sodium nitrite, and potassium sorbate. The HMA mice of both genotypes were randomly allocated to control, i.e., water supplied, and additive-treated groups. The size of each experimental group was 5 to 8 mice. The metagenomic analysis of faecal samples was performed using a QIIME2 software package.
Bioinformatic analysis revealed that the mixture of additives decreased the number of ASVs initially present in both wild-type and Nod2-/-mice. The decrease in alpha diversity of faecal samples as represented by Faith’s index was significant only for wild-type mice (Figure 2
A). However, it is essential to note that the number of ASVs in control Nod2-/- mice (n
= 88) was lower when compared to control wild-type mice (n
= 94), and the additive treatment decreased the average number of ASVs in both genotypes to 70 ASVs. Therefore, the effect of additive treatment on alpha diversity in Nod2-/- mice was partly limited by an already decreased number of ASVs due to the genotype effect.
Analysis of beta diversity confirmed a significant effect of genotype on the composition of gut microbiota. The finding that Nod2 is a major factor in the regulation of commensal microbiota is not novel and was published in cooperation with our group a long time ago [27
]. However, this project focused primarily on the possible dysbiosis-inducing capacity of additives in two genetic backgrounds, i.e., wild-type and Nod2-deficient. In this regard, the beta diversity analysis demonstrates that human gut microbiota in both wild-type and Nod2-/- settings is susceptible to additives. Remarkably, the gut microbiota in Nod2-/- environment (p
= 0.001) was significantly more susceptible to additives compared to the microbiota in a wild-type environment (p
= 0.005). This finding could be explained by impaired defensin production in Nod2-/- mice which renders their microbiota more susceptible to environmental triggers. The tight clustering of samples in all four experimental groups and the separation of control and additive-treated samples are convincingly documented by the PCoA plot (Figure 3
The taxonomic analysis of the samples from wild-type mice showed that the effect of additives was mostly represented by a decrease in the relative abundance of Firmicutes and an increase in Verrucomicrobia and Proteobacteria. The abundance of Bacteroidetes was almost unaffected. The analysis of the additive-treated samples from Nod2-/- mice also showed a decrease in Firmicutes and an increase in Bacteroidetes and Proteobacteria. Therefore, the common feature in both wild-type and Nod2-/- samples is a depletion of Firmicutes, mostly from the Clostridiales order, and the expansion of Proteobacteria, mostly from the Burkholderiales order.
The differential abundance analysis gives a more dynamic image of the effect of additives on gut microbiota than simple taxonomic analysis. For example, in wild-type mice, the seemingly unharmed Bacteroidales
are actually affected (Figure 6
). Six ASVs assigned to the Bacteroidaceae
family are depleted, and simultaneously four ASVs assigned to the same family are expanded, resulting in an unaltered relative abundance at the phylum or order level (Table 1
). Moreover, the total depletion of the Clostridiales
order is compounded of a decrease of 23 ASVs and an increase of four ASVs (Figure 6
and Table 1
). The same is true for Nod2-/- mice, where the expansion of Bacteroidetes
consists of an increase of 10 ASVs assigned to the Bacteroidaceae
family, and at the same time, a decrease of 4 ASVs assigned to the same family, and the depletion of Clostridiales
consists of an increase of 16 ASVs and a decrease of 35 ASVs (Figure 5
and Table 1
). Notably, the expansion of Proteobacteria
in Nod2-/- mice consisted solely of an increase of three ASVs with no Proteobacteria
Therefore, it is important to note that not all ASVs in particular taxonomic levels and categories are susceptible or resistant to the effect of additives, and the overall abundance as represented by the taxonomic bar plot (Figure 5
) consists of depleted and expanded ASVs.
Published research on the effects of antimicrobial food additives on human gut microbiota, i.e., preservatives, is scarce. Irwin et al. [14
] showed on four probiotic strains that sulphites may in vitro inhibit bacterial growth at concentrations regarded as safe. Another group showed that antimicrobial biopolymer ε-polylysine could trigger transient compositional alterations in the mouse gut microbiome [15
]. Our recent in vitro research on human gut isolates demonstrates that some gut microbes are highly susceptible to commonly used antimicrobial food additives. For example, the growth of Bacteroides coprocola
is inhibited by 50% in the presence of only 30 ng ml-1 of sodium nitrite. Other highly susceptible strains include Clostridium tyrobutyricum
and Lactobacillus paracasei.
Alarmingly, some combinations of additives, such as a benzoate-nitrite-sorbate mixture, showed a very high degree of synergism [16
]. However, the research on non-antimicrobial additives presented some notable findings recently. Chassaing et al. [12
] showed that food emulsifiers alter human gut microbiota composition and may promote colitis and metabolic syndrome, and Rodriguez-Palacios et al. [13
] reported that the artificial sweetener Splenda can induce Proteobacteria
The limitations of the work are as follows: firstly, the gut microbiota isolated from other donors, i.e., dissimilar as to gender, age, or race, might exhibit different levels of susceptibility to additives. Ideally, the metagenomic analysis should be performed on a higher number of donor samples, but the upscaling of this work was not possible due to a limited project budget. Secondly, the gut microbiota acquired from a human donor and transplanted into germ-free mice does not represent an identical copy of the original microbiota due to several factors including the transient exposure to oxygen during transfer or different luminal conditions, such as pH, bile acids, IgA specificity, and defensins. Finally, the gut microbiota isolated from a European donor already exhibits a reduced diversity—the most sensitive strains have already been lost—due to the exposure to antibiotics, medications, and food additives. Therefore, we speculate that the gut microbiota isolated from donors living in less-developed rural areas may be even more susceptible to additive exposure.
Moreover, it is important to note that the observed effects of additives on human gut microbiota might be caused by a single additive as well as any combination of additives. Therefore, further experiments testing each additive separately are needed. Most importantly, further experiments should focus on the extensive analysis of direct or indirect, via microbiota, effects of additives on host physiology including gut permeability, immune system parameters, histology, serum parameters, and the bioavailability of additives.
To sum up the findings of this project, we conclude that the mixture of commonly used antimicrobial food additives at the exposure levels reached in EU populations has the potential to induce human gut microbiota dysbiosis characterised by a decrease in microbiota diversity, a depletion of the Clostridiales order, and the expansion of Proteobacteria phylum and, that the Nod2-/- genotype is particularly susceptible to gut microbiota disruption. These findings are highly important because this type of dysbiosis is associated with many immune-mediated and metabolic disorders.