The impact of food additives, artificial sweeteners and domestic hygiene products on the human gut microbiome and its fibre fermentation capacity

A great amount of research has investigated the role of dietary nutrients, or dietary patterns in general, on the gut microbiome. Dietary fibre has attracted the most interest, mainly due to the inability of the human body to utilise it, and the capability of the gut microbiome to ferment it using a broad spectrum of enzymes not encoded in the human genome cannot encode [1]. Short-chain fatty acids (SCFA) are the end-product of fibre fermentation and the SCFA produced are dependent on the host’s diet and microbiome composition. Species within Bacteroides produce primarily acetic acid and propionic acid [2, 3]; members of Clostridium leptum cluster produce butyric acid from fibre fermentation and Bifidobacterium produces lactate and acetic acid from carbohydrate fermentation [4]. The branch chain fatty acids (BCFA) iso-butyric acid and iso-valeric acid are produced from protein breakdown, particularly in the absence of fermentable carbohydrate. Yet, the human gut microenvironment dynamics are more complex and characterised by an extensive degree of inter-species synergy and cross-feeding. It is, therefore, important to study the interactions between diet and the gut microbiome in the context of the entire microbial community and not as microbes in isolation. SCFA are critical bacterial products involved, not only locally in gut health, but in whole-body homeostasis. Along with an increased microbial diversity, high butyric acid concentration in the gut has been used as an indicator of healthy status of the microbiome. In contrast, reduced diversity, low luminal production of SCFA and dysbiosis have been proposed as primary events of inflammatory bowel disease, diabetes and obesity [5‐8].

Our diet has evolved enormously and rapidly over the last century, in parallel with food preservation and processing and increased use of industrialised and domestic hygiene products. While food industrialisation has protected humanity from infectious diseases, the secondary effect this may have on gut microbiome-dependent host health, and the net impact on the incidence of non-communicable diseases has only relatively recently been considered. A Mediterranean diet with increased consumption of legumes, cereals, fruit and vegetables, and its health-promoting effects, influences the gut microbiome [9]. The Western diet, which includes food additives and preservatives, has contrastingly been associated with non-communicable diseases [10]. Food additives and artificial sweeteners have become increasingly prevalent within our diet, with more than 50% of available food in UK households being ultra-processed [11]. While food additives are evaluated rigorously for their effects on the host, health testing of food additives fails to include their effect on the human gut microbiome and by proxy long-term host health [12]. Recent studies in animals have indicated that food additives can have adverse effects on colonic and cardiovascular health, mediated by the gut microbiome and changes in the gut mucus layer. It has been shown that food emulsifiers, such as polysorbates and carboxymethyl cellulose can increase intestinal permeability, alter microbiota composition, promote Escherichia coli translocation across the epithelium and in M cells in-vitro causing gut inflammation [10, 13]. Likewise, the body of evidence on artificial sweeteners indicates that there are adverse metabolic outcomes in rodents owing to the onset of microbial dysbiosis [14‐16]. Cumulative ingestion of residual products from regular use of domestic hygiene products may influence the human gut microbiome and, by extension, the health of the host. In epidemiological research, increased use of dishwashers, which reduce residual domestic detergent on dishware and consequent accidental ingestion, was associated with a decrease in cardiovascular disease [17]. Although for some food additives, artificial sweeteners and domestic hygiene products a large amount will be digested or degraded in the upper part of the gastrointestinal tract, residual amounts can still reach the colon. Others, like carrageenans and carboxylmethyl cellulose will reach the colon in similar amounts to those ingested.

It has become increasingly accepted that a diverse gut microbiome with high production of SCFA, particularly butyric acid, is an independent biomarker of host health. It is also known that diet influences the gut microbiome structure and function, including its fibre fermentation capacity [32]. However, relatively little is known about what effect that food additives, artificial sweeteners and accidental exposure to domestic hygiene products might have on the gut microbiome. As our diet has become more industrialised and is expected to become even more so to sustain food availability, it is important to understand the beneficial or detrimental effect food additives may have on the gut microbiome, and by extension to host health, to guide current and future use.

This study measured the effect of thirteen commonly used food additives, artificial sweeteners, and domestic hygiene products on the healthy gut microbiome composition and its fermentation capacity using in-vitro human microbiome batch fermentations. Changes in the ability of the gut microbiome to ferment fibre and produce SCFA and quantitative changes in major bacterial groups were measured and the summary results of this study are presented in Fig. 7. In addition to these analyses, the global microbiome composition and community structure were characterised using 16S rRNA gene amplicon sequencing as displayed in Fig. 8. Six of the additives affected the production of SCFA, five influenced the global microbiome community structure and nine altered the concentration of dominant microbial groups. Only toothpaste, stevia and carboxymethyl cellulose showed no or minimal effects on the broad composition and fermentation capacity of the faecal microbiome. However, for the additives for which an effect was observed, changes in microbiome composition and SCFA concentrations varied considerably among them; in terms of both the microorganisms or SCFA affected as well as the direction of this effect. Thus, this study highlights that the gut microbiome is modifiable in different ways by different additives. These variable effects of various food additives also suggest that their impact on the gut microbiome needs to be studied separately for each, in combination with each other, and in addition to other macronutrients, micronutrients and fibre in our diet.

There is increasing interest in the effect of the food industrialisation on human health and particularly on non-communicable disease, such as inflammatory bowel disease and diabetes [10, 20]. In previous studies, these effects were associated directly or indirectly with the microbiome of the large bowel. A food additive can affect gut homeostasis by influencing either the gut microbiome, the mucus layer or both. Carrageenan-kappa, upon consumption, has been associated with an increased prevalence of intestinal lesions in animal models [33], highlighting a detrimental effect on the mucosal barrier. Recent evidence from experiments in mice shows that this effect may be mediated by changes in the abundance of Akkermansia muciniphila, a potent anti-inflammatory bacterium. The results of the current study show that similar effects were observed with inhibition in the growth of Bifidobacterium and B. coccoides cluster, members of which have beneficial effects for the host [34]. Similarly, the dietary emulsifiers carboxymethyl cellulose and polysorbate-80 have been proposed to directly alter human microbiome composition and ex-vivo gene expression, potentiating intestinal inflammation [21]. Although in our current study no effect of carboxymethyl cellulose was seen on the fermentation capacity or on shifts in major bacterial groups, polysorbate-80 decreased the growth of Bifidobacterium and C. leptum and the relative abundance of other Firmicutes as confirmed by the results of both qPCR and 16S rRNA gene sequencing. Inorganic sulphite salts are frequently used to stop fermentation in wine and beer as well as antioxidants in food. This bacteriostatic effect of sodium sulphite was observed for members of the genus Bifidobacterium and the cluster B. coccoides and this effect may give a growth advantage to E. coli and Bilophila wadsworthia which gain energy through sulphite respiration [35]. Irwin et al. have previously described the bactericidal effects of sodium sulphite on probiotic-type bacteria, common members of the human gut microbiome [36]. The exact opposite effects were observed for the growth of Bifidobacterium, the cluster B. coccoides and E. coli when either maltodextrin or the aspartame-based sweetener was present. This is likely to be because maltodextrin is an artificially produced glucose polymer which, if not absorbed in the small intestine, has prebiotic properties in the colon [37]. Therefore, the increase in the probiotic genus Bifidobacterium and B. coccoides and the corresponding decrease in E. coli is most likely due to the fact that the former two use maltodextrin for growth [38, 39] instigating fermentation, production of acetic acid and creating an acidic environment in which E. coli growth is suppressed. Interestingly, changes in the absolute concentration of these three dominant bacterial groups, quantified with qPCR, were not in parallel the absence of effects observed using next generation sequencing. Sodium benzoate has been shown to decrease plasma ammonium levels by reducing glycine metabolism to treat patients with urea-cycle-disorder and acute hyperammonaemia [40, 41]. Use of sodium benzoate in this study increased the beneficial Bifidobacterium but reduced E. coli and the concentration of BCFA, suggesting that protein fermentation and potentially production of ammonia from bacterial metabolism in the gut is diminished. Similar to maltodextrin, these effects were not observed with in-depth characterisation of the microbiome using 16S rRNA sequencing. However, discordant results are to be expected as qPCR provides an absolute quantification of broader groups of bacteria and 16S rRNA sequencing offers proportional representation of the overall microbial community. The aspartame-based sweetener we used in this study was rich in maltodextrin in addition to, aspartame and acesulflame potassium. This, therefore, prevented the study of aspartame in isolation. However, the absence of major differences between the maltodextrin and the aspartame-based sweetener suggests that most of the effect seen on the gut microbiome comes from maltodextrin with no major contributions of aspartame and acesulfame potassium; at least in the amount we tested in this experiment which equals 8% of the estimated daily intake which might carry-over to the gut.

Crohn’s disease has been characterised by a gut microbiome with a reduced number of Firmicutes, such as species belonging to C. leptum, Bifidobacterium and Bacteroidetes and an increase in Proteobacteria, particularly E. coli strains with adherent and invasive properties [22, 24]. Interestingly, the addition of cinnamaldehyde, a cinnamon ingredient, or dishwashing detergent increased the E. coli and decreased the C. leptum and B. coccoides growth. A similar effect was also observed for polysorbate-80 with a diminished abundance of butyrate-producing species and increase in a species of Bilophila, a hydrogen sulphide producer implicated in colitis in IL-10 knockout mice [42]. Assuming that the gut microbial dysbiosis seen in patients with Crohn’s disease is a primary defect of the disease, and such species are implicated in disease pathogenesis, these findings suggest that consumption of cinnamon-containing food, polysorbate-80 and accidental ingestion of residual detergent on crockery and utensils may exacerbate dysbiosis and influence disease outcomes. Dishwashing detergent contains surfactants, which lower the surface tension, potentially causing degradation of mucus layer and the mucosal barrier to break-down thus potentially affecting microbial composition [43] particularly microbes which are adjacent to epithelial cells. Many Firmicutes like Faecalibacterium and Subdoligranulum are butyric acid-producing bacteria; hence the diminishing production of butyric acid here coincides with the decline in the concentration and abundance of these taxa with the addition of cinnamaldehyde, and dishwashing detergent. Firmicutes constitute a large proportion of the bacteria in the human gut microbiome, therefore, a significant change to the composition and functionality found within this phylum could, in theory, have detrimental consequences to the host. Butyric acid, for example, is the preferable energy substrate for the colonocytes and regulates regulatory T cells which play an important role in cell-mediated immunity [44]. A similar effect on C. leptum was seen for polysorbate-80 and a modest one for titanium dioxide. Collectively this evidence proposes that these additives could exacerbate the microbial dysbiosis seen in inflammatory bowel disease.

This study looked at the effect of food additives, artificial sweeteners and domestic hygiene products on the gut microbiome composition and fibre fermentation capacity in healthy human individuals, using batch fermentations with human faecal inoculum; thus, complementing previous research in animals. Although in the current study the SCFA and microbiome composition profile of the control, following 24 h fermentation, are in accordance to those that occur in the human gut, batch fermentation is a snapshot and not an exact simulant of human gut physiology and its complex dynamics [45, 46]. This may explain some of the discrepancies between the findings of this study and previous research [10]. Batch faecal fermentations do, however, provide crucial preclinical data, under well-controlled experimental conditions. They enable exploration of various additives at the same time and the direct effect on the gut microbiome in isolation of the host effect; hence bridging the gap between animal research and human trials. The data generated from this study offer important insights on where future research on additives should be directed, using animal experiments and human randomised controlled trials. In our case, this may be relevant for cinnamaldehyde, polysorbate-80, sodium sulphite, sodium benzoate, sucralose and dishwashing detergent but not for carboxymethyl cellulose, and stevia. While maltodextrin and the aspartame-based sweetener influenced the gut microbiome composition and production of SCFA, they did not induce dysbiosis and their effect might be considered favourable by inhibiting the growth of E. coli, thus promoting Bifidobacterium and correspondingly increasing the production of acetic acid and propionic acid. This bifidogenic effect of maltodextrin, an artificial glucose polymer has been observed previously too [47]. These findings are in contrast to evidence suggesting that maltodextrin induces dysbiosis promoting gut inflammation [10]. Such discrepancies might be explained by broad differences in the methodology applied among studies and the fact that in the current study we explored the effect of maltodextrin on the gut microbiome in isolation of the host and gut physiology. However, maltodextrin is the main source of carbohydrate in proprietary feeds used for the amelioration of gut inflammation with exclusive enteral nutrition in active Crohn’s disease [48, 49]. This reproducible clinical evidence challenges our current perceptions on the role of maltodextrin on gut inflammation.

This study contributes to the limited knowledge on the effect of food additives, artificial sweeteners and domestic hygiene products on the human gut microbiome composition and fibre fermentation capacity. We have shown that the presence of certain additives changed the microbial composition, and this became similar to the gut microbiome seen in individuals with either inflammatory bowel disease or obesity. For other additives, their effects were counterintuitive and opposite to animal research, implicating them in gut inflammation, and by proxy to human inflammatory bowel disease [10, 50]. This study underpins the importance of evaluating each additive separately and not grouped by their functional class. Here, we lay the groundwork for future research into individual additives on the gut microbiome composition and its fermentation capacity measured over a longer time period both in public health research and in the context of therapeutic interventions in patients with established dysbiosis, including patients with inflammatory bowel disease.