Non-caloric artificial sweeteners (NAS) are generally considered a safe alternative to sugars for use as a food additive, especially for people controlling their sugar intake.
However, previous studies have given conflicting or inconclusive results regarding the effectiveness of NAS in aiding weight loss. In fact, some have concluded that NAS may actually cause weight gain, though they noted that further research was needed.
In 2014, a research group from the Weizmann Institute of Science demonstrated that consumption of some commonly used NAS can drive the development of glucose intolerance – an early hallmark of metabolic syndrome and type II diabetes – by inducing compositional and functional alterations in intestinal microbiota.
These results are important for both understanding the complex interactions and activities of the intestinal microbiota on metabolic processes, and understanding how non-caloric artificial sweeteners can impact health. With its widespread use in products for the purpose of helping consumers to improve their health, these findings show that it may be doing the very opposite.
Chronic NAS consumption exacerbates glucose intolerance in mice
Commercial formulations of saccharin, sucralose or aspartame were added to the drinking water of lean C57BI/6 mice fed on normal chow. Mice that drank only water, or solutions containing 100% glucose or sucrose were used as controls.
Saccharin was further investigated. C57BI/6 mice were fed a high-fat diet (HFD), while consuming 0.1mg/ml commercial saccharin or pure glucose (control). The doses of NAS corresponded to the FDA acceptable daily intake in humans, adjusted to mouse weights.
Results showed significant glucose intolerance in both lean and HFD-fed mice exposed to the artificial sweeteners, while the water, glucose and sucrose controls all showed comparable glucose tolerance curves.
Gut microbiota mediates NAS-induced glucose intolerance in mice
Mouse groups consuming NAS in both lean and HFD states were treated with a gram-negative-targeting broad-spectrum antibiotic to test whether the gut microbiota may regulate the observed NAS effects. The antibiotics were ciprofloxacin and metronidazole (“antibiotics A”). A second antibiotic treatment with gram-positive-targeting antibiotic vancomycin (“antibiotics B”) was also conducted.
After four weeks of antibiotic treatment, the differences in glucose intolerance between NAS-drinking mice and the controls were abolished in both the lean and HFD-fed groups for both treatments. These results suggest that NAS-induced glucose intolerance may be mediated by the intestinal microbiota.
Faecal transplantation experiments were conducted to establish a causal relationship between glucose intolerance and the microbiota. The microbiota of lean mice drinking commercial saccharin or glucose (control) was transferred into lean mice that have never encountered saccharin in their diets.
Recipients of the microbiota from saccharin-drinking mice exhibited an impaired glucose tolerance response compared to recipients of the microbiota from control. Transfers of the microbiota composition of HFD-fed mice drinking saccharin also resulted in the same glucose intolerance phenotype.
NAS mediate distinct functional alterations to the microbiota
The faecal microbiota composition of the various mouse groups were compared by sequencing their 16S ribosomal RNA gene.
Results showed that the mice that had been drinking saccharin had a distinct microbiota composition that clustered separately from both their starting microbiome, and from that of the control groups. Similarly, the microbiome of the recipients of saccharin-drinking mice showed a separate clustering compared to the recipients of the glucose-drinkers. Overall, a similar configuration of dysbiosis (bacterial imbalance) was observed in all mice groups that consumed saccharin compared to all control groups.
The functional changes of the microbiota after NAS exposure was studied by performing shotgun metagenomics sequencing on the faecal samples before saccharin consumption, and after eleven weeks of treatment. Comparisons of relative species abundance was done by mapping sequencing reads to the human microbiome project reference genome database.
The results showed that saccharin treatment induced large chances in microbial relative species abundance.
Mapping of metagenomic reads to a gut microbial gene catalogue also showed a change in the abundance of pathways in saccharin-consuming mice compared to their control counterparts. Over-represented pathways in saccharin-drinking mice included glycan degradation pathways, marking enhanced energy harvest which may be associated with obesity in mice and humans. This is due to fermentation of glycans into various compounds, including short-chain fatty acids (SCFAs), which may serve as precursors and/or signalling molecules for de novo synthesis of glucose or lipids in the host.
The increase in these pathways is attributable to reads originating from the same bacteria that was found to be abundant in saccharin-consuming mice. This was consistent with the sharp increase in the abundance of particular genera observes in 16S rRNA analysis of NAS-consuming mice.
Other pathways were also enriched in microbiomes of the NAS-consuming mics, including:
- Starch, sucrose, frustose and mannose metabolism
- Folate, glycerolipid and fatty acid biosynthesis
The glucose transport pathways were decreased by comparison.
In HFD-fed mice with saccharin intake, other pathways were also enriched, including:
- Ascorbate and aldarate metabolism – reportedly enriched in leptin-receptor-deficient diabetic mice
- Lipopolysaccharide biosynthesis – linked to metabolic endotoxaemia
- Bacterial chemotaxis – reportedly enriched in obese mice
NAS directly modulate the microbiota to induce glucose intolerance
The direct effects of saccharin on gut microbiota was observed by culturing faecal matter from naïve mice in the presence of saccharin (5mg/ml) or control growth media. After incubation, the cultures were administered in mice that hadn’t been exposed to saccharin before. In vitro stool culture with saccharine had an increase in the same bacterial phylums as that of the in vivo faecal samples. Transfers of the in vitro samples resulted in a significantly higher glucose intolerance, compared to the mice receiving the control culture.
Shotgun metagenomics sequencing analysis of the in vitro saccharin treatment also showed induction of similar functional alterations as the in vivo saccharin treatment. Glycan degradation pathways were highly enriched in both settings.
NAS in humans associate with impaired glucose tolerance
A human study was conducted by examining the relationship between long-term NAS consumption, based on validated food frequency questionnaire, and various clinical parameters in an ongoing clinical nutritional study.
There were significant positive correlations between NAS consumption and several symptoms of metabolic syndrome, including:
- Increased weight
- Increased waist-to-hip ratio
- Higher fasting blood glucose
- Increased glucose intolerance
- Glycosylated haemoglobin
The levels of glycosylated haemoglobin was significantly increased in a subgroup of individuals of NAS consumers to non-NAS consumers. Of this cohort, the 16S rRNA were characterised and a statistically positive correlation between multiple taxonomic entities and NAS consumption.
An initial assessment of whether the relationship was causal, seven healthy volunteers who did not normally consume NAS were followed for a week, during which they consumed the maximal acceptable daily intake of commercial saccharin as recommended by the FDA. Four of the seven individuals developed a significantly poorer glycaemic response after NAS consumption after around five days of NAS consumption. The other three did not feature improved glucose tolerance.
The microbiome configurations of the NAS responders were assessed by 16S rRNA analysis, and clustering of the microbiota appeared different to that of the non-responders, both before and after NAS consumption. The microbiomes from the non-responders also showed little change in composition during the week of study, while there were pronounced compositional changes observed in the NAS responders.
NAS exposure was then transferred from two NAS responders and two NAS non-responders into groups of normal-chow-fed mice. The transfer of post-NAS exposure stool from NAS responders induced a significant glucose intolerance in the recipients compared to the response from pre-NAS exposure stool from the same responders. The post-NAS exposure stool of the non-responders, however, did not induce abnormal glucose tolerance in the mice. The response was indistinguishable from that of pre-NSA exposure stools from the same individuals.