Invitro microbiota model recapitulates and predicts individualised sensitivity to dietary emulsifier

Héloïse Rytter; Sabrine Naimi; Gary Wu; Jim Lewis; Maeva Duquesnoy; Lucile Vigué; Olivier Tenaillon; Eugeni Belda; Marta Vazquez-Gomez; Nina Touly; Djésia Arnone; Fuhua Hao; Ruth E Ley; Karine Clément; Laurent Peyrin-Biroulet; Andrew D Patterson; Andrew T Gewirtz; Benoit Chassaing

doi:10.1136/gutjnl-2024-333925

. 2025 Jan 27;74(5):e333925. doi: 10.1136/gutjnl-2024-333925

Invitro microbiota model recapitulates and predicts individualised sensitivity to dietary emulsifier

Héloïse Rytter ^1,², Sabrine Naimi ², Gary Wu ³, Jim Lewis ⁴, Maeva Duquesnoy ^1,², Lucile Vigué ⁵, Olivier Tenaillon ⁵, Eugeni Belda ^6,⁷, Marta Vazquez-Gomez ^6,⁷, Nina Touly ^8,^9,^10,¹¹, Djésia Arnone ^8,^9,^10,¹¹, Fuhua Hao ¹², Ruth E Ley ¹³, Karine Clément ^6,¹⁴, Laurent Peyrin-Biroulet ^8,^9,^10,^11,¹⁵, Andrew D Patterson ¹², Andrew T Gewirtz ¹⁶, Benoit Chassaing ^1,^2,^11,^✉

¹Microbiome-Host Interactions, INSERM U1306, CNRS UMR6047, Institut Pasteur, Université Paris Cité, Paris, France

²INSERM U1016, CNRS UMR8104, Mucosal Microbiota in Chronic Inflammatory Diseases, Université de Paris, Paris, France

³Division of Gastroenterology, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA

⁴Center for Clinical Epidemiology and Biostatistics, Division of Gastroenterology and Hepatology,Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA

⁵Robustness and Evolvability of Life, CNRS UMR10 8104, INSERM U1016, Université Paris Cité, Paris, France

⁶Inserm, Nutrition and Obesities: Systemic Approaches Research Unit, NutriOmics, Sorbonne University, Paris, France

⁷Unité de Modélisation Mathématique et Informatique des Systèmes Complexes, UMMISCO, IRD, Sorbonne Université, Bondy, France

⁸INFINY Institute, Nancy University Hospital, Vandœuvre-lès-Nancy, France

⁹INSERM, NGERE, University of Lorraine, Nancy, France

¹⁰FHU-CURE, Nancy University Hospital, Vandoeuvre les Nancy, France

¹¹CHRU Nancy, IHU Infiny, Vandoeuvre-les-Nancy, France

¹²Center for Molecular Toxicology and Carcinogenesis, Veterinary and Biomedical Sciences, The Pennsylvania State University, University Park, Pennsylvania, USA

¹³Microbiome Science, Max-Planck-Institute for Biology, Tübingen, Germany

¹⁴Assistance Publique Hôpitaux de Paris, Nutrition Department, Pitié-Salpêtrière Hospital, Paris, France

¹⁵Department of Gastroenterology, Nancy University Hospital, Vandoeuvre-lès-Nancy, France

¹⁶Center for Inflammation, Immunity and Infection, Digestive Disease Research Group, Department of Biology, Georgia State University, Atlanta, Georgia, USA

Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Additional supplemental material is published online only. To view, please visit the journal online (https://doi.org/10.1136/gutjnl-2024-333925).

LP-B received consulting and lecture fees as well as a grant from Takeda. None of the other authors declare competing interests.

^✉

Dr Benoit Chassaing; benoit.chassaing@pasteur.fr

PMCID: PMC12013555 PMID: 39870396

Abstract

Background

Non-absorbed dietary emulsifiers, including carboxymethylcellulose (CMC), directly disturb intestinal microbiota, thereby promoting chronic intestinal inflammation in mice. A randomised controlled-feeding study (Functional Research on Emulsifiers in Humans, FRESH) found that CMC also detrimentally impacts intestinal microbiota in some, but not all, healthy individuals.

Objectives

This study aimed to establish an approach for predicting an individual’s sensitivity to dietary emulsifiers via their baseline microbiota.

Design

We evaluated the ability of an in vitro microbiota model (MiniBioReactor Arrray, MBRA) to reproduce and predict an individual donor’s sensitivity to emulsifiers. Metagenomes were analysed to identify signatures of emulsifier sensitivity.

Results

Exposure of human microbiotas, maintained in the MBRA, to CMC recapitulated the differential CMC sensitivity previously observed in FRESH subjects. Furthermore, select FRESH control subjects (ie, not fed CMC) had microbiotas that were highly perturbed by CMC exposure in the MBRA model. CMC-induced microbiota perturbability was associated with a baseline metagenomic signature, suggesting the possibility of using one’s metagenome to predict sensitivity to dietary emulsifiers. Transplant of human microbiotas that the MBRA model deemed CMC-sensitive, but not those deemed insensitive, into IL-10^−/− germfree mice resulted in overt colitis following CMC feeding.

Conclusion

These results suggest that an individual’s sensitivity to emulsifier is a consequence of, and can thus be predicted by, examining their baseline microbiota, paving the way to microbiota-based personalised nutrition.

Keywords: INTESTINAL MICROBIOLOGY, INFLAMMATORY BOWEL DISEASE, MUCUS, NUTRITION

WHAT IS ALREADY KNOWN ON THIS TOPIC

Commonly used dietary emulsifiers can perturb intestinal microbiota, fostering chronic intestinal inflammation and metabolic deregulations.
A high level of interindividual variations has been observed, in human, regarding emulsifier-induced microbiota perturbations.

WHAT THIS STUDY ADDS

Sensitivity to dietary emulsifiers of a given faecal sample can be predicted using in vitro microbiota modelling system.
Dietary emulsifier sensitivity associated with a metagenomic signature commonly observed in healthy individuals as well as in patients with chronic inflammatory conditions.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

These findings highlight the need to carefully consider interindividual variations in emulsifier sensitivity status for personalised nutrition.

Introduction

The intestinal microbiota has major beneficial impacts on host physiology, including driving immune system maturation and impeding colonisation of pathogens.^{1 2} However, detrimental alterations in microbiota composition and/or function, referred as dysbiosis, can promote chronic diseases with an inflammatory component, including inflammatory bowel disease (IBD) and various inter-related metabolic deregulations such as obesity and diabetes.³,⁶ Myriad environmental factors can detrimentally impact the intestinal microbiota, especially the consumption of highly processed foods, thus contributing to the pathogenesis of these disease conditions.^{7 8} A common feature of processed food products is the presence of various food additives, such as emulsifiers, which improve texture and/or extend shelf-life.⁹ Dietary emulsifiers have long been considered non-toxic compounds, in part, because they pass through the intestinal tract and are largely eliminated in faeces without having been absorbed. However, such restriction to the intestinal lumen positions them to directly interact with the intestinal microbiota. Indeed, recent evidence suggests that detrimental impacts of emulsifiers on the intestinal microbiota may be linked to increased incidence of several chronic inflammatory diseases.^{10 11} Mechanistically, these compounds are thought to act by promoting microbiota encroachment into the normally sterile inner mucus layer,¹²,¹⁴ positioning bacteria to promote chronic intestinal inflammation that, depending on host genetics, can manifest as colitis or metabolic deregulations.^{15 16}

The degree to which emulsifiers impact humans was recently explored via a randomised double-blind controlled-feeding study focusing on the synthetic emulsifier carboxymethylcellulose (CMC). We found that CMC consumption induces alteration in microbiota composition and faecal metabolome in healthy participants.¹⁷ However, the extent of microbiota perturbation was not uniform. Rather, select subjects exhibited profound CMC-induced alterations in microbiota composition, accompanied by microbiota encroachment, suggesting high interindividual variability in CMC sensitivity.¹⁷ Such individualised sensitivity to CMC was likely a consequence of baseline microbiota in that transplantation of microbiotas from individuals presenting strong CMC-induced microbiota alterations into germfree IL10^−/− mice resulted in severe colitis following exposure to CMC.¹⁸ In stark contrast, transplant of microbiotas from CMC-resistant individuals conferred protection against CMC-induced pathologies,¹⁸ highlighting the central role played by the microbiota in driving detrimental responses to emulsifiers, as well as the strong microbiota-driven interindividual variations in emulsifier sensitivity. Consequently, there is a need for a practical means to identify emulsifier-sensitive individuals, and thus alerting them that they would likely benefit from avoiding these additives.

That mice receiving microbiota transplants acquire the CMC sensitivity status of the human faecal donor demonstrates that the information needed to determine CMC sensitivity is contained within an individual’s baseline microbiota. However, use of germfree mice is not a practical or scalable means of determining sensitivity to food additives. Hence, we here explored the use the MiniBioReactor Arrray (MBRA) microbiota model and metagenome-based bioinformatic modelling to predict the CMC sensitivity of a given microbiota. We found the MBRA model recapitulated the CMC sensitivity status observed in the Functional Research on Emulsifiers in Humans (FRESH) trial. Furthermore, study of faeces from FRESH control subjects in the MBRA model, and subsequently IL-10^−/− germfree mice, indicated the MBRA could predict an individual’s CMC sensitivity status. We also found that CMC sensitivity was associated with a metagenomic signature, suggesting that emulsifier sensitivity may ultimately be reliably predictable from one’s metagenome. These results support development of microbiome-based personalised nutrition strategies.

Results

MBRA model faithfully recapitulated microbiotas of the FRESH subjects

The results of our recently reported randomised double-blind controlled-feeding study of non-metabolisable emulsifier CMC, entitled Functional Research on Emulsifiers in Humans (FRESH),¹⁷ accorded with findings in mice that CMC impacts the intestinal microbiota. However, the extent of such impacts was heterogeneous in that only two of the seven subjects fed this additive exhibited stark changes in microbiota composition and microbiota encroachment in response to this additive. Such subjects were deemed CMC-sensitive. Here, we sought to develop a means of identifying CMC-sensitive microbiotas without in vivo studies. We employed the MiniBioReactor Arrays (MBRA),^{19 20} an in vitro microbiota modelling system that allows dynamic and stable culture of human-derived microbiota under anaerobic conditions (online supplemental figure S1A). Faecal samples from the 16 FRESH subjects (9 participants from the control group and 7 participants from the CMC-exposed groups) were used to inoculate triplicate MBRA chambers (figure 1A). MBRA chambers were then sampled longitudinally (eight times over 8 days), assaying microbiota composition via Illumina-based 16S rRNA gene sequencing.

We first analysed the extent to which the MBRA system could reproduce the well-appreciated interindividual heterogeneity in microbiota composition. Data were displayed as a principal coordinates analysis (PCoA) plot wherein each colour corresponds to an individual subject. This approach revealed strong per-donor clustering (figure 1B), indicating that the MBRA model was capable of stably cultivating donor-specific microbiotas for up to 12 days. Moreover, taxonomic analysis of the stabilised MBRA microbial communities showed that the MBRA captured and maintained donor specificity at the order and phylum levels (figure 1C and online supplemental figure S2A). At the phylum level, Firmicutes was the most abundant MBRA microbiota member, averaging 39%, followed by Bacteroidetes (35%) and Proteobacteria (21%) (online supplemental figure S2A). Interindividual differences were more apparent at the order level. For example, Verrucomicrobiales ranged from 0.009% to 7.647%, mostly driven by Akkermansia muciniphila, which promotes intestinal and metabolic health.²¹,²³ We also observed donor-dependent variations in the relative abundance of Faecalibacterium prausnitzii, ranging from 0.00% to 11.00%, a bacterium purported to protect against chronic intestinal inflammation.²⁴,²⁶ Both of these bacteria are strict anaerobes and are often lost in in vitro-based microbiota culturing. Thus, we viewed their presence in the MBRA as supporting its fastidiousness. We next assessed the extent to which the stabilised MBRA microbiomes corresponded to the FRESH subjects from which they were derived. PCoA plotting revealed that each donor microbiome clustered squarely within the MBRA time point samples derived from it (online supplemental figure S2B). Moreover, Bray-Curtis distances between each baseline MBRA sample and all baseline donor samples indicated that baseline MBRA microbiomes were far more similar to their cognate donor than the other donors (online supplemental figure S2C). Comparison of bacterial composition of the poststabilisation MBRA chambers with baseline faeces composition showed the ability of MBRA to reproduce donor specificity was most apparent at the order levels (online supplemental figure S3). Thus, MBRA faithfully reproduced the interindividual variations in microbiota composition of the FRESH cohort.

The extent to which MBRA maintained donor-specific microbiota metabolomes was investigated by targeted ¹H NMR-based metabolomic analysis. Analysis of data by PCoA of the Bray-Curtis distance found a high level of per-participant metabolome clustering (time point 72 hours, Permanova p <0001, online supplemental figure S4A). Moreover, we observed a strong interindividual variation in the concentration of various metabolites, such as short-chain fatty acids, lactate and tryptophan (online supplemental figure S4B). Altogether, these data support the suitability of the MBRA to study donor-specific microbiotas ex vivo.

MBRA reproduced the individualised CMC-sensitivity status observed in the FRESH study

The FRESH study found that sensitivity to CMC was not uniform. Specifically, as shown here in figure 1D, we reported that two of the seven CMC-fed subjects exhibited marked alterations in their microbiota composition over the 10-day course of their CMC consumption.¹⁷ That these two individuals also developed microbiota encroachment suggests the compositional changes were functionally important.¹⁷ We refer to these two individuals, shown in orange in figure 1D, as being CMC-sensitive and deem the other five CMC-fed subjects, shown in blue, to be relatively CMC-insensitive individuals. Here, we investigated if such CMC-sensitivity could be recapitulated by the MBRA model. MBRA were inoculated with baseline faecal samples from each FRESH subject (three replicate MBRA per faecal sample) and allowed an 8-day stabilisation period. MBRA microbiotas were then left untreated for eight additional days or administered CMC (0.1% w/v) for 5 days, followed by a post-treatment phase of 3 days (online supplemental figure S1B). MBRA samples were collected longitudinally (online supplemental figure S1B) and subjected to microbiome analysis (online supplemental figure S1C). Bacterial density and Shannon diversity were stable for all donors irrespective of CMC exposure (online supplemental figure S5A,B). However, temporal Bray-Curtis dissimilarity analysis revealed stark donor-specific differences that corresponded to the CMC-sensitivity previously observed in the FRESH study. Specifically, MBRA corresponding to both CMC-sensitive subjects displayed stark CMC-induced changes. In contrast, MBRA that had been inoculated with faeces from the CMC-insensitive subjects remained relatively stable in response to CMC (figure 1E and online supplemental figure S6A). These results indicated that the MBRA had reproduced interindividual variations in CMC-sensitivity observed in the FRESH study.

Prediction of CMC sensitivity status via metagenomic signature and MBRA model

The CMC sensitivity of subjects in the control arm of the FRESH study is, by definition, unknown, but by extrapolation one might expect it would contain 2–3 CMC-sensitive subjects. Hence, we used metagenomic analysis combined with MBRA approach to investigate this notion.

We first performed shotgun metagenomic analysis on faecal samples collected at baseline from the FRESH subjects (figure 2A). Data from the CMC group were analysed at the imputed UniRef90 level using MaAsLin2, comparing CMC-sensitive and CMC-insensitive subjects. Such analysis revealed 217 markers significantly associated with CMC-sensitivity status (q<0.05, figure 2B). We next analysed the abundance of these markers in the FRESH control subjects. Microbiomes from two of the nine FRESH control subjects displayed a pattern of expression of these markers that was highly similar to that of CMC-sensitive subjects, as assessed by PCoA plots and heatmap representation (figure 2C,D), suggesting that these two individuals harboured CMC-sensitive microbiotas.

We next tested the metagenomic marker-based prediction of CMC sensitivity status via the MBRA model. Specifically, MBRA chambers were inoculated with faecal samples from the nine FRESH control subjects and subsequently treated, or not, with CMC. Analysis of such MBRA microbiotas validated the metagenomic marker-based predictions in that the two faecal microbiotas predicted to be CMC- sensitive, but not the other seven, were significantly perturbed by CMC treatment as reflected by increases in Bray-Curtis dissimilarity distances (figure 2E and online supplemental figure S6B). These two CMC-sensitive microbiotas did not stand out from the others in terms of Shannon diversity nor bacterial density at baseline or following CMC exposure (online supplemental figure S5C,D). Moreover, examining previously generated 16S sequencing data from the FRESH study did not reveal any of these control subjects to have microbiomes that were inherently unstable over the course of the FRESH study (figure 2F).

That our metagenomic marker-based predictions matched the MBRA results for all nine individuals from the FRESH control group by chance was unlikely (less than 1%). Nonetheless, we felt the need to further interrogate our approach. Hence, we assessed whether MaAsLin2 comparisons not based on known CMC sensitivity might also, somehow, predict CMC sensitivity. For this, we applied the MaAsLin2 algorithm to metagenomes of arbitrarily selected groups (2 vs 4, not based on the sensitivity status) among the CMC arm of the FRESH study. All possible combinations were computed, and none of them were found to correlate with the CMC sensitivity status observed in the MBRA. Thus, training a prediction algorithm with metagenomes from CMC-sensitive subjects, but not arbitrary metagenomes, predicted a microbiota’s sensitivity to CMC in the MBRA model.

In vivo validation of CMC-sensitivity predictions

We next investigated the extent to which microbiotas that the metagenomic signature and MBRA model identified as being CMC- sensitive were truly capable of conferring a biologically significant consequence of CMC exposure in vivo. Faecal microbiotas from two FRESH control subjects deemed CMC-sensitive by MBRA, as well as two deemed insensitive, were transplanted into colitis-prone IL-10^−/- germfree mice (n=2 donors per condition and recipient mice per donor). After a 2-week period of microbiota stabilisation, mice were, or were not, administered CMC via their drinking water (1% w/v) for 16 weeks (figure 3A), at which time intestinal inflammation was assessed. Il-10^−/− mice harbouring microbiotas corresponding to CMC-insensitive MBRA lacked signs of intestinal inflammation following CMC consumption, as evidenced by unchanged histopathological colonic score and stable colonic infiltration with CD68⁺ macrophages (figure 3B–E). In stark contrast, Il-10^−/− mice administered microbiotas corresponding to CMC-sensitive MBRA developed robust intestinal inflammation on CMC consumption, as evidenced by significant colon shortening, increased histopathological scores mostly driven by mucosal damages, and increased CD68⁺ macrophage infiltration (figure 3B–E and online supplemental figure S7). To identify potential key bacterial taxa associated with intestinal inflammation, correlation coefficients were calculated between inflammatory scores and bacterial abundance at the specie level. This approach identified bacteria, including Adlercreutzia equolifaciens and Frisingicoccus caecimuris, that positively correlated with intestinal inflammation (online supplemental figure S8). Future work will be needed to decipher the role played by these microbiota members in emulsifier-driven inflammation.

One cardinal driver of an array of inflammatory diseases is encroachment of microbiota into the normally near-sterile inner mucus region.1213 15 18 27,²⁹ Microbiota encroachment has been observed in mice consuming emulsifiers,^{12 13 30} as well as in select FRESH participants consuming CMC.¹⁷ Hence, we examined microbiota localisation via confocal microscopy in water-treated or CMC-treated IL-10^−/− mice colonised with microbiota from either predicted CMC-sensitive or predicted CMC-insensitive FRESH participants. In agreement with the above measures of inflammation, CMC consumption had no impact on microbiota-epithelium distance in IL-10^−/− mice colonised with microbiota predicted to be CMC-insensitive (figure 3F,G). In stark contrast, CMC consumption drove severe microbiota encroachment in IL-10^−/− mice that had been colonised with both CMC-sensitive microbiotas tested. Specifically, microbiota-epithelium distance decreased in response to CMC by 14,81 µm and 12,04 µm compared with water-treated groups (figure 3F,G). Collectively, these results support the validity of the MBRA model as a means to identify CMC-sensitive microbiomes.

CMC sensitivity associated with a baseline microbiota signature within the whole FRESH cohort

The identification of two additional CMC-sensitive microbiotas, and seven additional CMC-insensitive microbiotas, proved an opportunity to further investigate microbiota-mediated CMC sensitivity. An array of genomic and functional microbiota parameters was compared between the 4 CMC-sensitive and 12 CMC-insensitive donors that were identified by FRESH and/or MBRA studies (figure 4A). First, we measured, using MBRA-derived samples, the ability of the various microbiota to directly degrade/consume CMC. We found that, irrespective of donor, CMC concentrations in the MBRA effluent were only slightly decreased relative to the administered dose (online supplemental figure S9A,B). This result aligns with the view that CMC traverses the intestine without being metabolised and argues against its differential metabolism contributing to one’s CMC sensitivity status. Next, we investigated the possibility that emulsifier sensitivity status reflects more general alterations in microbiota-derived metabolomes. For this purpose, we performed targeted metabolomic analysis employing ¹H NMR-based assay on MBRA samples collected in the middle of the CMC treatment phase. While this approach revealed subtle CMC-induced variations, for example, in methanol level, such alterations were independent of CMC sensitivity status (online supplemental figure S10).

We next further probed the potential of microbiomes to predict CMC sensitivity. Neither PCoA analysis of the Bray-Curtis distances using MetaPhlAn4-generated taxonomical data, nor taxonomical barplots, at phylum and order levels, revealed differential clustering based on the CMC sensitivity status (online supplemental figure S11A–C). Moreover, MaAsLin2³¹ found that none of the 461 species that were identified in these samples significantly differed in abundance between CMC-sensitive and CMC-insensitive participants. These data suggest that taxonomical composition of a given faecal sample may not be sufficient to predict emulsifier sensitivity status.

We next analysed shotgun metagenomic data at the UniRef90 level, using the HUMAnN 3.0 pipeline.³² PCoA analysis of the Bray-Curtis distances comparing the 1 649 550 obtained metagenomic markers did not reveal CMC sensitivity-based clustering (figure 4B). However, MaAsLin2-based analysis revealed a specific CMC-sensitivity metagenomic signature of 78 genomic markers whose abundance strongly associated with CMC sensitivity status (q<0.05, figure 4C,D). The algorithm used by MaAsLin2 uses a statistical correction to correct for the high number of variables being analysed. Nonetheless, we worried that deep comparisons of metagenomes from any two groups of subjects might reveal analogous sets of distinguishing genomic markers. Hence, we applied the MaAsLin2 algorithm to metagenomes of an arbitrarily selected group, namely male versus female participants (n=7 and 9, respectively) and found 0 distinguishing metagenomic markers. We also applied this approach to the subjects that had previously been randomised to be in the CMC versus the control arm of the study (7 vs 9), as well as to randomly selected group sizes that matched the number of CMC-sensitive versus insensitive subjects (4 vs 12). Again, in both cases, we found 0 metagenomic markers that distinguished these groups. Such lack of association in these arbitrarily and randomly selected groups supports the notion that these 78 metagenomic markers are indeed associated with emulsifier sensitivity status. Interestingly, all 78 metagenomic markers displayed increased abundance in CMC-sensitive microbiotas, suggesting that sensitivity status is driven by the presence, rather than absence, of select metabolic pathways. Most of these metagenomic markers were from Clostridium, Dorea, Coprobacillus and Escherichia genera (online supplemental figure S12A). Analysis of these markers at the protein level indicated that the majority of them encoded uncharacterised functions (online supplemental figure S12B), and future studies will be needed to decipher the exact role played by these metagenomic markers in driving sensitivity towards dietary emulsifiers. Nonetheless, these results further support the potential of metagenomic analysis as means of marking the CMC-sensitivity status of an individual’s microbiota.

Prevalence of metagenomic markers of emulsifier sensitivity in independent cohorts of healthy participants and patients

We next applied the CMC sensitivity metagenomic signature, identified in the Philadelphia PA, USA FRESH cohort, to European cohorts (figure 5A). Specifically, we defined the prevalence of each of the 78 metagenomic markers in subjects in the Metacardis cohort, which includes both healthy participants (N=458) and obese patients with (N=657) or without type 2 diabetes (T2D) (N=430).³³ We also applied this approach to Crohn’s disease patients (N=199) present in Nancy Hospital’s Mucosa cohort. Both of these cohorts exhibited a heterogeneous prevalence of the CMC sensitivity genomic markers independent of health status (figure 5B–D) although, as expected, a high level of interindividual variation was observed. Specifically, individual subjects harboured 0–66 (of 78) metagenomic markers, while the prevalence of each marker varied between 1% and 88%. Future studies will be needed to investigate the specificity and sensitivity of this signature in identifying CMC sensitivity in various cohorts.

Discussion

The increased prevalence of an array of chronic inflammatory diseases, particularly as world food supplies increasingly industrialise, suggests that components of modern diets may be environmental (ie, non-genetic) disease determinants. An example of this notion is dietary emulsifiers, a ubiquitous class of food additives, which, in mice, promotes inflammation that can manifest in a variety of forms depending on the genetics of the host. Emulsifiers promote/exacerbate colitis in an array of mouse models, primarily by altering intestinal microbiota composition and/or gene expression.^{12 13} That these findings are conceptually relevant to humans is supported by observations that ‘exclusion diets’, which lack food additives, including emulsifiers, show efficacy in maintaining remission in paediatric IBD.³⁴,³⁶ Furthermore, the FRESH study, which directly examined the impacts of the synthetic emulsifier CMC on healthy humans, supported the notion that this additive alters the intestinal microbiota. However, the extent of CMC’s impact was highly variable, suggesting that only some individuals are CMC-sensitive. Mechanism that drove interindividual variations in CMC sensitivity, as well as whether this parameter can be predicted without exposing a host to this additive, remained to be elucidated.

We, here, report that the MBRA in vitro microbiota modelling system, previously shown to maintain key features of human faecal microbial communities,^{20 37 38} reproduced the clinically observed intermicrobiota variations in emulsifier sensitivity observed in the FRESH study. Further, MBRA studies suggested that some subjects in the control arm of the FRESH study were also CMC-sensitive. These microbiotas were confirmed to indeed be CMC-sensitive via faecal transplant studies. Specifically, germfree IL-10^−/− mice transplanted with faeces from microbiota MBRA studies deemed CMC-sensitive, developed severe colitis following CMC feeding while similarly treated recipient microbiotas MBRA deemed CMC-insensitive displayed only mild inflammation. Thus in vitro microbiota modelling enables a means to predict whether a given microbiota would be prone to promoting inflammation on exposure to CMC without performing any in vivo studies.

Our results suggest that prediction of CMC sensitivity may also be predictable from an individual’s metagenome. Indeed, we found that training the MaAsLin2 algorithm with metagenomes of the CMC arm of the FRESH study enabled it to correctly stratify the CMC sensitivity status of the nine FRESH control subjects. However, using the resulting refined signature generated from the full FRESH cohort to predict CMC sensitivity in new cohorts proved less straightforward. Specifically, training MaAsLin2 with the metagenomes of the 4 CMC-sensitive and 12 CMC-insensitive FRESH subjects identified 78 functional markers as being positively correlated with emulsifier sensitivity. While these markers could be found in European cohorts, the frequency at which they were present resulted in poor ability to stratify CMC-sensitivity status. Such inapplicability of the FRESH cohort-derived signature to distinct cohorts may reflect the need for more extensive sets of training data and/or that some predictive microbiome signatures are inherently cohort-specific.

Analysis of the metagenomic signature as means of imputing mechanisms that mediate CMC sensitivity did not prove informative. Many of these markers (47%) coded for unidentified function while others were assigned to general functional categories, for example, 15% were related to transcription. Thus, further research is needed to determine how these markers might mediate CMC sensitivity. In any event, that all identified markers were positively associated with CMC sensitivity suggests that this phenotype is driven by the presence of select microbiome genes rather than the absence of factors that provide emulsifier resistance. Lastly, the identification of sensitivity markers at the functional level, but not at the taxonomical levels, suggests that a given—or association of—microbiota member(s) is not sufficient to drive emulsifier deleterious effects in all susceptible individuals. Accordingly, presence of these metagenomic markers was heterogeneous in independent cohorts in which we’d expect some individuals would be CMC-sensitive.

The concept that selects components of diet selectively impact disease is not restricted to IBD, but rather can be extended to numerous other chronic inflammatory diseases, especially the large subset of such diseases in which gut microbiota composition is thought to be a disease determinant. Indeed, in the case of emulsifiers, epidemiological studies have recently highlighted the association between dietary emulsifier consumption and an increased risk of various cancer,³⁹ T2D,⁴⁰ as well as cardiovascular diseases.⁴¹ Moreover, we previously reported that the impact of various food additives on microbiota is compound specific,¹⁴ indicating that CMC is not the only food additive able of inducing deleterious effect on health. Our general hypothesis holds that the extent to which a given individual to develop deleterious response following chronic exposure to a particular food additive will likely be dependent on its intestinal microbiota.

The food industry has commonly added specific compounds, such as artificial sweeteners and stabilisers, to meet societal demands at lower costs. Several studies have reported their effects on human health and their interactions with the microbiota.⁴²,⁴⁴ Moreover, a recently published clinical study suggests that non-nutritive sweeteners may induce glycaemic alterations that are person-specific and dependent on the composition of the microbiome.⁴⁵ Our recent findings, along with other studies using in vitro microbiota systems, highlight the potential of MBRA to predict sensitivity to such compounds.^{14 37 46} Further analyses are needed to elucidate correlations between sensitivity to specific dietary compounds and microbiota composition, enabling the prediction of an individual’s response based on microbiota profiling. In conclusion, we report the use of in vitro microbiota modelling and sequence-based microbiome analysis to predict whether an individual’s intestinal ecosystem will be perturbed, and consequently their proneness to disease enhanced, by consumption of a specific food additive. Extending this approach to larger cohorts and additional food additives will advance implementation of personalised nutritional approaches potentially enabling individuals to avoid specific food additives that might pose a danger to them.

Materials and methods

Reagents

Sodium CMC (E466, average MW~250 000) was purchased from (Sigma, St. Louis, Missouri, USA).

Faecal sample collection

Fresh faecal samples were collected from participants of a randomised, controlled-feeding study that took place at the University of Pennsylvania’s Center for Human Phenomic Science, as described previously¹⁷ (figure 2A and ClinicalTrials.gov, number NCT03440229). The study design and the experimental protocol were previously detailed in.¹⁷ Stool samples were collected without preservatives or stabilisers at the time of inclusion, and aliquoted and frozen at −80°C.

Setup and use of the in vitro MBRA system and treatment with CMC

The MBRA system was prepared and used as described previously.^{19 20} Briefly, the MBRA system is housed in an anaerobic chamber and consists, per system, of 24 individual chambers, as presented online supplemental figure S1A. MBRA chambers were held on a magnetic stand for continual homogenisation and were connected to two 24-channel peristaltic pumps with low flow rate capabilities (205S peristaltic pump with 24-channel drive, Watson-Marlow). After autoclaving, the prepared MBRA system was placed in an anaerobic chamber for at least 72 hours. MBRA chambers were then filled with 15 mL of Bioreactor Medium (BRM) and subsequently inoculated with freshly prepared faecal slurry. For this, faecal samples were resuspended at 10 % w/v in anaerobic phosphate buffered saline (D‐PBS) (Gibco–Life Technologies) within an anaerobic environment, vortexed for 5 min and centrifuged at 800 rpm for 5 min at 20°C. Supernatant was subsequently within an anaerobic environment and filtered through a 100 μm filter to remove any residual particles. Finally, 3.8 mL of this faecal slurry was used to inoculate each MBRA chambers, with six independent chambers inoculated per donor. Twenty-four hours after inoculation, constant flow was initiated at 1.875 mL/hour, corresponding to an 8-hour retention time. After a 72-hour microbiota stabilisation phase, obtained chambers were treated, in triplicates for each donor, with BRM medium containing CMC at 0.1% (w/v, added prior autoclaving) or control sterile BRM medium. At the 120-hour time point, CMC treatment was discontinued by connecting all chambers to emulsifier-free BRM sterile medium. Samples were collected longitudinally (400 μL), as presented online supplemental figure S1B, with CMC treatment occurring between 0 hour and 120 hours postinoculation. Samples were then stored at −80°C until analysis (online supplemental figure S1C).

Bacterial DNA extraction

As described previously,¹⁴ DNA was extracted from frozen MBRA suspension (50 μL) using a QIAamp 96 PowerFecal QIAcube HT kit (Qiagen Laboratories) with mechanical disruption (Qiagen TissueLyser II). Briefly, 650 µL of prewarmed buffer PW1 was added to 50 µL of each sample. Samples were thoroughly homogenised using bead-beating with a TissueLyser before centrifuging the plate at 4000 rpm for 5 min at 20°C in order to pellet beads and particles. 400 µL of supernatant was added into a new 96-well plate containing 150 µL of Buffer C3. After mixing and incubation on ice for 5 min, centrifugation was performed at 4000 rpm for 5 min at 20°C. 300 µL of each supernatant was added to a new 96-well S-block plate, and 20 µL of Proteinase K was added and incubated for 10 min at room temperature. The following steps were next performed on a QIAcube high-throughput robot: addition of 500 µL of Buffer C4, DNA binding to a QIAamp 96 plate, column wash using AW1 (800 µL), AW2 (600 µL) and ethanol (400 µL) and DNA elution using ATE buffer (100 µL).

Bacterial density quantification through 16S rRNA qPCR

Extracted DNAs were diluted 1/10 with sterile DNA-free water and subjected to quantitative PCR using the 16S V4 specific primers 515F 5’-GTGYCAGCMGCCGCGGTAA-3’ and 806R 5’-GGACTACNVGGGTWTCTAAT-3’ on a LightCycler 480 (Roche) using QuantiFast SYBR Green PCR Kit (Qiagen). Data were analysed by The LightCycler 480 Software (Roch molecular system), and results were expressed as relative values.

Microbiota analysis by 16S rRNA gene sequencing using Illumina technology

16S rRNA gene amplification and sequencing were performed using the Illumina MiSeq technology following the protocol described previously.^{47 48} The 16S rRNA genes, region V4, were PCR amplified from each sample using a composite forward primer and a reverse primer containing a unique 12-base barcode, designed using the Golay error-correcting scheme, which was used to tag PCR products from respective samples.⁴⁷ The forward primer 515F was used: 5’-AATGATACGGCGACCACCGAGATCTACACGCTXXXXXXXXXXXXTATGGTAA TTGTGTGYCAGCMGCCGCGGTAA-3’: the italicised sequence is the 5’ Illumina adapter, the 12 X sequence is the Golay barcode, the bold sequence is the primer pad, the italicised and bold sequence is the primer linker and the underlined sequence is the conserved bacterial primer 515F. The reverse primer 806R used was 5’-CAAGCAGAAGACGGCATACGAGATAGTCAGCCAGCCGGACTACNVGGGTWTCTAAT-3’: the italicised sequence is the 3’ reverse complement sequence of Illumina adapter, the bold sequence is the primer pad, the italicised and bold sequence is the primer linker and the underlined sequence is the conserved bacterial primer 806R. PCR reactions consisted of 5PRIME HotMasterMix (Quantabio, Beverly, Massachusetts, USA), 0.2 µM of each primer, 10–100 ng template and reaction conditions were 3 min at 95°C, followed by 30 cycles of 45 s at 95°C, 60 s at 50°C and 90 s at 72°C on a Biorad thermocycler. PCRs products were visualised by gel electrophoresis and purified with Ampure magnetic purification beads (Agencourt, Brea, California, USA). Products were then quantified (Quant-iT PicoGreen dsDNA assay), and a master DNA pool was generated from the purified products in equimolar ratios. The pooled products were quantified using the Quant-iT PicoGreen dsDNA assay and sequenced using an Illumina MiSeq sequencer (paired-end reads, 2×250 bp) at the Genom’IC sequencing platform from Institut Cochin, Paris, France.

16S rRNA gene sequences analysis

16S rRNA sequences were analysed using QIIME2—version 2022.⁴⁹ Sequences were demultiplexed and quality filtered using Dada2 method⁵⁰ with QIIME2 default parameters in order to detect and correct Illumina amplicon sequence data, and a table of Qiime 2 artefact was generated. A tree was next generated, using the align-to-tree-mafft-fasttree command, for phylogenetic diversity analyses, and alpha and beta diversity analyses were computed using the core-metrics-phylogenetic command. For taxonomy analysis, features were assigned to operational taxonomic units with a 99% threshold of pairwise identity to the Greengenes reference database V.13.8.⁵¹ Unprocessed sequencing data are deposited in the European Nucleotide Archive under accession number PRJEB84075.

¹H NMR-based metabolomic analysis

MBRA samples collected during the pretreatment phase (−24 hours time point), during the treatment phase (72 hours time point) and during the post-treatment phase (192 hours time point) were used for metabolomic analysis. Samples preparation for NMR were performed as previously described.⁵² ¹H NMR spectra were acquired on a Bruker Avance NEO 600 MHz spectrometer equipped with an inverse cryogenic probe (Bruker Biospin, Germany) at 298 K. A typical 1D NMR spectrum named NOESYPR1D was acquired for each sample. Metabolites were assigned and confirmed with a series of 2D NMR spectra, as described previously.⁵³ All 1H NMR spectra were adjusted for phase and baseline using Chenomx. The chemical shift of 1H NMR spectra were referenced to sodium 3-trimethylsilyl [2,2,3,3-d4] propionate (TSP) at δ 0.00. The relative contents of metabolites were calculated by normalising to the total sum of the spectral integrals. The quantification of metabolites, including CMC, was calculated by NMR peak area against TSP using Chenomx.

Microbiota analysis by shotgun metagenomic using Illumina technology

DNAs were extracted from faecal samples of the 16 FRESH donors used in the MBRA system using the QIAamp Fast DNA Stool Mini Kit (Qiagen). Extracted DNAs were fragmented by sonication (six cycles of 90 s) using a bath sonicator (Bioruptor Plus sonication device, Diagenode) at 4°C. Library preparation was performed using the Invitrogen Collibri PS DNA Library Prep Kit for Illumina (ThermoFisher) according to manufacturer’s recommendations. The purified library was then subjected to sequencing using an Illumina NextSeq500 (paired-end reads, 2×250 bp) at the Genom’IC sequencing platform from Institut Cochin, Paris, France. Obtained sequences were quality filtered and sequencing adapters were removed from the resulting sequences via cutadapt.⁵⁴ These quality-filtered reads were then grouped via MetaPhlAn V.2.0⁵⁵ into taxonomical categories and via HUMAnN3³² into functional categories.

The metagenomes from two following additional cohorts were used:

The MUCOSA cohort (Development of a Non-Invasive Mucosal Healing Index in Crohn’s Disease), from Nancy, Besançon and Saint-Étienne, France and includes 199 patients with Crohn’s disease, aged 18–75 years.
The MetaCardis cohort (http://www.metacardis.net/) from various European countries with a total of 1545 participants, including 458 healthy individuals, 430 obese individuals without T2D and 657 obese individuals with T2D. Recruitment for the MetaCardis cohort occurred between 2013 and 2015 across several clinical institutions in Europe: Pitié-Salpêtrière Hospital, the Center of Research for Clinical Nutrition and the Institute of Cardiometabolism and Nutrition in France; the Integrated Research and Treatment Center (IFB) for Adiposity Diseases in Leipzig, Germany; and the Novo Nordisk Foundation Center for Basic Metabolic Research in Copenhagen, Denmark.

Mice experiments

Germ-free C57BL/6 IL-10^−/− male mice (C57BL/6NTac-Il10^em8Tac; Taconic model GF-16006) were ordered at 4–6 weeks of age and immediately subjected to faecal microbiota transplantation. For this purpose, mice were orally administered with 400–500 µL of the faecal suspension from CMC-predicted insensitive or CMC-predicted sensitive FRESH participants collected at the time of inclusion (eight mice per donor). Animals were maintained in isolated ventilated caging system (Isocages from Techniplast, West Chester, Pennsylvania, USA, 1–2 mice per cage) to protect them from environmental contamination⁵⁶ and in a controlled environment (12 hours day/night cycle, lights off at 7:00PM, 21°C±2°C, 42°C±13% of humidity) at Institut Cochin (Paris, France). Experimentations were under institutionally approved protocol (APAFIS#24 788-2019102806256593 v8) and mice were provided unlimited access to autoclaved food (Purina Rodent Chow #5021) and autoclaved water (Cristaline) ad libitum. After 2 weeks of acclimatisation post microbiota transplantation, mice received either autoclaved water (Cristaline) or autoclaved CMC solution (average MW~250 000, Sigma, St. Louis, Missouri, USA) at 1% diluted in water for 16 weeks, resulting in eight groups of mice (n=4, online supplemental figure S2). Water and CMC solution were autoclaved and changed every other week. Body weights over time were recorded, and fresh faeces samples were collected at various time points. After 16 weeks of treatment, mice were euthanised and periepididymal fat pad, spleen and caecum were weighed, and colon length was measured. Colon samples were collected for further analysis.

Histopathological analysis of colonic tissue and immunochemistry

Mouse proximal colons were placed in methanol-Carnoy’s fixative solution (60% methanol, 30% chloroform, 10% glacial acetic acid) for a minimum of 3 hours at room temperature and stored at 4°C. Tissues were then washed in methanol 2×30 min, absolute ethanol 2×15 min, ethanol/xylene (1:1) 15 min and xylene 2×15 min, followed by embedding in Paraffin with a vertical orientation and tissues were then sectioned at 6 μm thickness. Slides were stained with H&E using standard protocols for histological scoring, which was blindly determined on each colon as previously described.^{57 58} Briefly, each colon was assigned four scores based on the degree of epithelial damage and inflammatory infiltrate in the mucosa, submucosa and muscularis/serosa.⁵⁷ Each of the four scores was multiplied by a coefficient 1 if the change was focal, 2 if it was patchy and 3 if it was diffuse 25 and the 4 individual scores per colon were added. For immunochemistry, stainings were performed on the automaton Leica Bond RX. Slides were unmasked at pH6 and incubated 30 min with an anti-CD68 antibody (abcam, EPR23917-164, 462 ab283654) 1:500 diluted and washed. The revelation system (‘Bond Polymere Refine’ kit, DS9800, Leica) included a secondary antibody HRP conjugated and revealed by DAB. High-resolution images were finally acquired using a Lamina Slide Scanner (Perkin Elmer) at the Hist’IM platform (INSERM U1016, Institut Cochin, Paris, France). Cell-based quantification of CD68+interstitial cells was performed by counting. Four images per mice were analysed, and positively immune-stained cells were counted in a blinded manner, using the ImageJ V.1.45 software and the ‘cell-counter’ plug-in.

Immunostaining of mucins and localisation of bacteria by fluorescent in situ hybridisation

Mucus immunostaining was paired with fluorescent in situ hybridisation, as previously described,⁵⁹ in order to analyse bacteria localisation at the surface of the intestinal mucosa. Briefly, colonic human biopsies (approximately 15 cm from the anal verge, correlating approximately with the rectosigmoid junction) and mouse tissues (proximal colon, 2nd cm from the cecum) containing faecal material were placed in methanol-Carnoy’s fixative solution (60% methanol, 30% chloroform and 10% glacial acetic acid) for a minimum of 3 hours at room temperature and stored at 4°C. Tissues were then washed in methanol 2×30 min, absolute ethanol 2×15 min, ethanol/xylene (1:1) 15 min and xylene 2×15 479 min, followed by embedding in Paraffin with a vertical orientation. Five-to-six mm sections were performed and dewaxed by xylene 60°C for 10 min, xylene for 10 min and 99.5% ethanol for 5 min. Hybridisation step was performed at 50°C overnight with EUB338 probe (50-GCTGCCTCCCGTAGGAGT-30, with a 5’ labelling using Alexa 647) diluted to a final concentration of 10 mg/mL in hybridisation buffer (20 mM Tris–HCl, pH 7.4, 0.9 M NaCl, 0.1% SDS, 20% formamide). After washing 10 min in wash buffer (20 mM Tris–HCl, pH 7.4, 0.9 M NaCl) and a quick wash in PBS, slides were incubated in block solution (5% fetal bovine serum in PBS) in darkness at 4°C for 30 min. Slides were then gently dried and PAP pen (Sigma-Aldrich) was used to mark around the section. Mucin-2 primary antibody (rabbit MUC2 antibody (C3), C-term, Genetex, GTX100664) was diluted 1:100 in block solution and applied overnight at 4°C. After washing 3×10 min in PBS, block solution containing anti-rabbit Alexa 488 secondary antibody diluted 1:300, Phalloidin-Tetramethylrhodamine B isothiocyanate (Sigma-Aldrich) at 1 mg/mL and Hoechst 33 258 (Sigma-Aldrich) at 10 mg/mL was applied to the section for 2 hours. After washing 3×10 min in PBS, the slides were mounted using Prolong anti-fade mounting media (Life Technolog×ies, Carlsbad, California, USA) and kept in the dark at 4°C. Images, observations and measurement of the distance between bacteria and epithelial cell monolayer were performed with a Spinning Disk IXplore using the Olympus cellSens imaging software 421 (V.2.3) at a frame size of 2048×2048 with 16-bit depth. A 405 nm laser was used to excite the 422 Hoechst stain (epithelial DNA), 488 nm for Alexa Fluor 488 (mucus), 488 nm for TRITC (actin), 423 and 640 nm for Alexa Fluor 647 (bacteria). Samples were imaged with a 20× objective immersion objective.

Statistical analysis

Significance was determined using t-tests, Mann-Whitney test, one-way analysis of variance (ANOVA) corrected for multiple comparisons with a Bonferroni post-test, two-way ANOVA corrected for multiple comparisons with a Bonferroni, Šidák or Dunnett post-test (or mixed-effect analysis when some values were missing). Differences were noted as significant if p≤0.05.

Patient and public involvement

Patients were not actively involved in this research.

Supplementary material

online supplemental figure 1

gutjnl-74-5-s001.pdf^{(7.7MB, pdf)}

DOI: 10.1136/gutjnl-2024-333925

online supplemental file 1

gutjnl-74-5-s002.pdf^{(843KB, pdf)}

DOI: 10.1136/gutjnl-2024-333925

Acknowledgements

The authors thank the Hist’IM and the Genom’IC platforms (INSERM U1016, Paris, France) for their help. Schematics were created with Biorender. (https://BioRender.com/w84e441).

No funders had any role in the design of the study and data collection, analysis, and interpretation, nor in manuscript writing.

Footnotes

Funding: This work was supported by a Starting Grant (grant agreement Invaders No. ERC-2018-StG-804135) and a Consolidator Grant (grant agreement InterBiome No. ERC-2024-CoG-101170920) from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program, ANR grants EMULBIONT (ANR-21-CE15-0042-01) and DREAM (ANR-20-PAMR-0002), grant for the AFA Crohn RCH France and the national program ’Microbiote’ from INSERM. This work was also supported by the French government through the France 2030 investment plan managed by the National Research Agency (ANR), as part of the ANR 23 IAHU 0012. Héloïse Rytter is supported by a fellowship from Paris Région—DIM One Health-DOH 2.0.

Provenance and peer review: Not commissioned; externally peer reviewed.

Patient consent for publication: Not applicable.

Ethics approval: Mouse experiments were approved and conducted in accordance with the guidelines of the local ethics committee for animal care at the Institut Cochin. All procedures adhered to the institutionally approved protocol (APAFIS#24788-2019102806256593 v8). This study also involved human faecal samples obtained from a previous study approved by the Institutional Review Board of the University of Pennsylvania (IRB#828422). The MUCOSA and MetaCardis cohort studies were approved by the local ethics committees at their respective institutions. All participants provided informed consent prior to their participation in the study.

Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

References

1.Hooper LV, Macpherson AJ. Immune adaptations that maintain homeostasis with the intestinal microbiota. Nat Rev Immunol. 2010;10:159–69. doi: 10.1038/nri2710. [DOI] [PubMed] [Google Scholar]
2.Cerf-Bensussan N, Gaboriau-Routhiau V. The immune system and the gut microbiota: friends or foes? Nat Rev Immunol. 2010;10:735–44. doi: 10.1038/nri2850. [DOI] [PubMed] [Google Scholar]
3.Dalal SR, Chang EB. The microbial basis of inflammatory bowel diseases. J Clin Invest. 2014;124:4190–6. doi: 10.1172/JCI72330. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Upadhyay KG, Desai DC, Ashavaid TF, et al. Microbiome and metabolome in inflammatory bowel disease. J Gastroenterol Hepatol. 2023;38:34–43. doi: 10.1111/jgh.16043. [DOI] [PubMed] [Google Scholar]
5.Hajj Hussein I, Dosh L, Al Qassab M, et al. Highlights on two decades with microbiota and inflammatory bowel disease from etiology to therapy. Transpl Immunol. 2023;78:101835. doi: 10.1016/j.trim.2023.101835. [DOI] [PubMed] [Google Scholar]
6.Ng SC, Shi HY, Hamidi N, et al. Worldwide incidence and prevalence of inflammatory bowel disease in the 21st century: a systematic review of population-based studies. Lancet. 2017;390:2769–78. doi: 10.1016/S0140-6736(17)32448-0. [DOI] [PubMed] [Google Scholar]
7.Chen J, Wellens J, Kalla R, et al. Intake of Ultra-processed Foods Is Associated with an Increased Risk of Crohn’s Disease: A Cross-sectional and Prospective Analysis of 187 154 Participants in the UK Biobank. J Crohns Colitis. 2023;17:535–52. doi: 10.1093/ecco-jcc/jjac167. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Sabino J, Lewis JD, Colombel J-F. Treating Inflammatory Bowel Disease With Diet: A Taste Test. Gastroenterology. 2019;157:295–7. doi: 10.1053/j.gastro.2019.06.027. [DOI] [PubMed] [Google Scholar]
9.Halmos EP, Mack A, Gibson PR. Review article: emulsifiers in the food supply and implications for gastrointestinal disease. Aliment Pharmacol Ther. 2019;49:41–50. doi: 10.1111/apt.15045. [DOI] [PubMed] [Google Scholar]
10.Marion-Letellier R, Amamou A, Savoye G, et al. Inflammatory Bowel Diseases and Food Additives: To Add Fuel on the Flames! Nutrients. 2019;11:1111. doi: 10.3390/nu11051111. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Martínez Steele E, Juul F, Neri D, et al. Dietary share of ultra-processed foods and metabolic syndrome in the US adult population. Prev Med. 2019;125:40–8. doi: 10.1016/j.ypmed.2019.05.004. [DOI] [PubMed] [Google Scholar]
12.Chassaing B, Koren O, Goodrich JK, et al. Dietary emulsifiers impact the mouse gut microbiota promoting colitis and metabolic syndrome. Nature New Biol. 2015;519:92–6. doi: 10.1038/nature14232. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Chassaing B, Van de Wiele T, De Bodt J, et al. Dietary emulsifiers directly alter human microbiota composition and gene expression ex vivo potentiating intestinal inflammation. Gut. 2017;66:1414–27. doi: 10.1136/gutjnl-2016-313099. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Naimi S, Viennois E, Gewirtz AT, et al. Direct impact of commonly used dietary emulsifiers on human gut microbiota. Microbiome. 2021;9:66. doi: 10.1186/s40168-020-00996-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Chassaing B, Raja SM, Lewis JD, et al. Colonic Microbiota Encroachment Correlates With Dysglycemia in Humans. Cell Mol Gastroenterol Hepatol. 2017;4:205–21. doi: 10.1016/j.jcmgh.2017.04.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Winter SE, Thiennimitr P, Nuccio S-P, et al. Contribution of flagellin pattern recognition to intestinal inflammation during Salmonella enterica serotype typhimurium infection. Infect Immun. 2009;77:1904–16. doi: 10.1128/IAI.01341-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Chassaing B, Compher C, Bonhomme B, et al. Randomized Controlled-Feeding Study of Dietary Emulsifier Carboxymethylcellulose Reveals Detrimental Impacts on the Gut Microbiota and Metabolome. Gastroenterology. 2022;162:743–56. doi: 10.1053/j.gastro.2021.11.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Daniel N, Wu GD, Walters W, et al. Human Intestinal Microbiome Determines Individualized Inflammatory Response to Dietary Emulsifier Carboxymethylcellulose Consumption. Cell Mol Gastroenterol Hepatol. 2024;17:315–8. doi: 10.1016/j.jcmgh.2023.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Auchtung JM, Robinson CD, Britton RA. Cultivation of stable, reproducible microbial communities from different fecal donors using minibioreactor arrays (MBRAs) Microbiome. 2015;3:42. doi: 10.1186/s40168-015-0106-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Auchtung JM, Robinson CD, Farrell K, et al. MiniBioReactor Arrays (MBRAs) as a Tool for Studying C. difficile Physiology in the Presence of a Complex Community. Methods Mol Biol. 2016;1476:235–58. doi: 10.1007/978-1-4939-6361-4_18. [DOI] [PubMed] [Google Scholar]
21.Schneeberger M, Everard A, Gómez-Valadés AG, et al. Akkermansia muciniphila inversely correlates with the onset of inflammation, altered adipose tissue metabolism and metabolic disorders during obesity in mice. Sci Rep. 2015;5:16643. doi: 10.1038/srep16643. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Dao MC, Everard A, Aron-Wisnewsky J, et al. Akkermansia muciniphila and improved metabolic health during a dietary intervention in obesity: relationship with gut microbiome richness and ecology. Gut. 2016;65:426–36. doi: 10.1136/gutjnl-2014-308778. [DOI] [PubMed] [Google Scholar]
23.Cani PD, Depommier C, Derrien M, et al. Akkermansia muciniphila: paradigm for next-generation beneficial microorganisms. Nat Rev Gastroenterol Hepatol. 2022;19:625–37. doi: 10.1038/s41575-022-00631-9. [DOI] [PubMed] [Google Scholar]
24.Martín R, Miquel S, Chain F, et al. Faecalibacterium prausnitzii prevents physiological damages in a chronic low-grade inflammation murine model. BMC Microbiol. 2015;15:67. doi: 10.1186/s12866-015-0400-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Cao Y, Shen J, Ran ZH. Association between Faecalibacterium prausnitzii Reduction and Inflammatory Bowel Disease: A Meta-Analysis and Systematic Review of the Literature. Gastroenterol Res Pract. 2014;2014:872725. doi: 10.1155/2014/872725. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Miquel S, Martín R, Rossi O, et al. Faecalibacterium prausnitzii and human intestinal health. Curr Opin Microbiol. 2013;16:255–61. doi: 10.1016/j.mib.2013.06.003. [DOI] [PubMed] [Google Scholar]
27.Kordahi M, Daniel N, Gewirtz AT, et al. Mucus-Penetrating Microbiota Drive Chronic Low-Grade Intestinal Inflammation and Metabolic Dysregulation. Gut Microbes. doi: 10.1080/19490976.2025.2455790. n.d. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Chassaing B, Ley RE, Gewirtz AT. Intestinal epithelial cell toll-like receptor 5 regulates the intestinal microbiota to prevent low-grade inflammation and metabolic syndrome in mice. Gastroenterology. 2014;147:1363–77. doi: 10.1053/j.gastro.2014.08.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Johansson MEV, Gustafsson JK, Holmén-Larsson J, et al. Bacteria penetrate the normally impenetrable inner colon mucus layer in both murine colitis models and patients with ulcerative colitis. Gut. 2014;63:281–91. doi: 10.1136/gutjnl-2012-303207. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Chassaing B, Gewirtz AT. Mice harboring pathobiont-free microbiota do not develop intestinal inflammation that normally results from an innate immune deficiency. PLoS One. 2018;13:e0195310. doi: 10.1371/journal.pone.0195310. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Mallick H, Rahnavard A, McIver LJ, et al. Multivariable association discovery in population-scale meta-omics studies. PLoS Comput Biol. 2021;17:e1009442. doi: 10.1371/journal.pcbi.1009442. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Beghini F, McIver LJ, Blanco-Míguez A, et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife . 2021;10:e65088. doi: 10.7554/eLife.65088. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Vieira-Silva S, Falony G, Belda E, et al. Statin therapy is associated with lower prevalence of gut microbiota dysbiosis. Nature New Biol. 2020;581:310–5. doi: 10.1038/s41586-020-2269-x. [DOI] [PubMed] [Google Scholar]
34.Levine A, Wine E, Assa A, et al. Crohn’s Disease Exclusion Diet Plus Partial Enteral Nutrition Induces Sustained Remission in a Randomized Controlled Trial. Gastroenterology. 2019;157:440–50. doi: 10.1053/j.gastro.2019.04.021. [DOI] [PubMed] [Google Scholar]
35.Sigall Boneh R, Van Limbergen J, Wine E, et al. Dietary Therapies Induce Rapid Response and Remission in Pediatric Patients With Active Crohn’s Disease. Clin Gastroenterol Hepatol. 2021;19:752–9. doi: 10.1016/j.cgh.2020.04.006. [DOI] [PubMed] [Google Scholar]
36.Sigall-Boneh R, Pfeffer-Gik T, Segal I, et al. Partial enteral nutrition with a Crohn’s disease exclusion diet is effective for induction of remission in children and young adults with Crohn’s disease. Inflamm Bowel Dis. 2014;20:1353–60. doi: 10.1097/MIB.0000000000000110. [DOI] [PubMed] [Google Scholar]
37.Hobson CA, Vigue L, Naimi S, et al. MiniBioReactor Array (MBRA) in vitro gut model: a reliable system to study microbiota-dependent response to antibiotic treatment. JAC Antimicrob Resist . 2022;4:dlac077. doi: 10.1093/jacamr/dlac077. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Hobson CA, Vigué L, Magnan M, et al. A Microbiota-Dependent Response to Anticancer Treatment in an In Vitro Human Microbiota Model: A Pilot Study With Hydroxycarbamide and Daunorubicin. Front Cell Infect Microbiol. 2022;12:886447. doi: 10.3389/fcimb.2022.886447. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Sellem L, Srour B, Javaux G, et al. Food additive emulsifiers and cancer risk: Results from the French prospective NutriNet-Santé cohort. PLoS Med. 2024;21:e1004338. doi: 10.1371/journal.pmed.1004338. [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Salame C, Javaux G, Sellem L, et al. Food additive emulsifiers and the risk of type 2 diabetes: analysis of data from the NutriNet-Santé prospective cohort study. Lancet Diabetes Endocrinol. 2024;12:339–49. doi: 10.1016/S2213-8587(24)00086-X. [DOI] [PubMed] [Google Scholar]
41.Sellem L, Srour B, Javaux G, et al. Food additive emulsifiers and risk of cardiovascular disease in the NutriNet-Santé cohort: prospective cohort study. BMJ. 2023;382:e076058. doi: 10.1136/bmj-2023-076058. [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Iizuka K. Is the Use of Artificial Sweeteners Beneficial for Patients with Diabetes Mellitus? The Advantages and Disadvantages of Artificial Sweeteners. Nutrients. 2022;14:4446. doi: 10.3390/nu14214446. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Markus V. Artificial sweetener-induced dysbiosis and associated molecular signatures. Biochem Biophys Res Commun. 2024;735:150798. doi: 10.1016/j.bbrc.2024.150798. [DOI] [PubMed] [Google Scholar]
44.Palmnäs MSA, Cowan TE, Bomhof MR, et al. Low-dose aspartame consumption differentially affects gut microbiota-host metabolic interactions in the diet-induced obese rat. PLoS One. 2014;9:e109841. doi: 10.1371/journal.pone.0109841. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Suez J, Cohen Y, Valdés-Mas R, et al. Personalized microbiome-driven effects of non-nutritive sweeteners on human glucose tolerance. Cell. 2022;185:3307–28. doi: 10.1016/j.cell.2022.07.016. [DOI] [PubMed] [Google Scholar]
46.Bonazzi E, Bretin A, Vigué L, et al. Individualized microbiotas dictate the impact of dietary fiber on colitis sensitivity. Microbiome. 2024;12:5. doi: 10.1186/s40168-023-01724-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Caporaso JG, Lauber CL, Walters WA, et al. Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. ISME J. 2012;6:1621–4. doi: 10.1038/ismej.2012.8. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Gilbert JA, Meyer F, Jansson J, et al. The Earth Microbiome Project: Meeting report of the “1st EMP meeting on sample selection and acquisition” at Argonne National Laboratory October 6th 2010. Stand Genomic Sci. 2010;3:249–53. doi: 10.4056/aigs.1443528. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Bolyen E, Rideout JR, Dillon MR, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019;37:852–7. doi: 10.1038/s41587-019-0209-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Callahan BJ, McMurdie PJ, Rosen MJ, et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods. 2016;13:581–3. doi: 10.1038/nmeth.3869. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.McDonald D, Price MN, Goodrich J, et al. An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J. 2012;6:610–8. doi: 10.1038/ismej.2011.139. [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Jiang L, Huang J, Wang Y, et al. Eliminating the dication-induced intersample chemical-shift variations for NMR-based biofluid metabonomic analysis. Analyst. 2012;137:4209–19. doi: 10.1039/c2an35392j. [DOI] [PubMed] [Google Scholar]
53.Tian Y, Zhang L, Wang Y, et al. Age-related topographical metabolic signatures for the rat gastrointestinal contents. J Proteome Res. 2012;11:1397–411. doi: 10.1021/pr2011507. [DOI] [PubMed] [Google Scholar]
54.Kechin A, Boyarskikh U, Kel A, et al. cutPrimers: A New Tool for Accurate Cutting of Primers from Reads of Targeted Next Generation Sequencing. J Comput Biol. 2017;24:1138–43. doi: 10.1089/cmb.2017.0096. [DOI] [PubMed] [Google Scholar]
55.Inam Z, Felton E, Burrell A, et al. Impact of Antibiotics on the Lung Microbiome and Lung Function in Children With Cystic Fibrosis 1 Year After Hospitalization for an Initial Pulmonary Exacerbation. Open Forum Infect Dis. 2022;9:ofac466. doi: 10.1093/ofid/ofac466. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Hecht G, Bar-Nathan C, Milite G, et al. A simple cage-autonomous method for the maintenance of the barrier status of germ-free mice during experimentation. Lab Anim . 2014;48:292–7. doi: 10.1177/0023677214544728. [DOI] [PubMed] [Google Scholar]
57.Katakura K, Lee J, Rachmilewitz D, et al. Toll-like receptor 9-induced type I IFN protects mice from experimental colitis. J Clin Invest. 2005;115:695–702. doi: 10.1172/JCI22996. [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Chassaing B, Srinivasan G, Delgado MA, et al. Fecal lipocalin 2, a sensitive and broadly dynamic non-invasive biomarker for intestinal inflammation. PLoS One. 2012;7:e44328. doi: 10.1371/journal.pone.0044328. [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Johansson MEV, Hansson GC. Preservation of mucus in histological sections, immunostaining of mucins in fixed tissue, and localization of bacteria with FISH. Methods Mol Biol. 2012;842:229–35. doi: 10.1007/978-1-61779-513-8_13. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

online supplemental figure 1

gutjnl-74-5-s001.pdf^{(7.7MB, pdf)}

DOI: 10.1136/gutjnl-2024-333925

online supplemental file 1

gutjnl-74-5-s002.pdf^{(843KB, pdf)}

DOI: 10.1136/gutjnl-2024-333925

Data Availability Statement

All data relevant to the study are included in the article or uploaded as supplementary information.

[R1] 1.Hooper LV, Macpherson AJ. Immune adaptations that maintain homeostasis with the intestinal microbiota. Nat Rev Immunol. 2010;10:159–69. doi: 10.1038/nri2710. [DOI] [PubMed] [Google Scholar]

[R2] 2.Cerf-Bensussan N, Gaboriau-Routhiau V. The immune system and the gut microbiota: friends or foes? Nat Rev Immunol. 2010;10:735–44. doi: 10.1038/nri2850. [DOI] [PubMed] [Google Scholar]

[R3] 3.Dalal SR, Chang EB. The microbial basis of inflammatory bowel diseases. J Clin Invest. 2014;124:4190–6. doi: 10.1172/JCI72330. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R4] 4.Upadhyay KG, Desai DC, Ashavaid TF, et al. Microbiome and metabolome in inflammatory bowel disease. J Gastroenterol Hepatol. 2023;38:34–43. doi: 10.1111/jgh.16043. [DOI] [PubMed] [Google Scholar]

[R5] 5.Hajj Hussein I, Dosh L, Al Qassab M, et al. Highlights on two decades with microbiota and inflammatory bowel disease from etiology to therapy. Transpl Immunol. 2023;78:101835. doi: 10.1016/j.trim.2023.101835. [DOI] [PubMed] [Google Scholar]

[R6] 6.Ng SC, Shi HY, Hamidi N, et al. Worldwide incidence and prevalence of inflammatory bowel disease in the 21st century: a systematic review of population-based studies. Lancet. 2017;390:2769–78. doi: 10.1016/S0140-6736(17)32448-0. [DOI] [PubMed] [Google Scholar]

[R7] 7.Chen J, Wellens J, Kalla R, et al. Intake of Ultra-processed Foods Is Associated with an Increased Risk of Crohn’s Disease: A Cross-sectional and Prospective Analysis of 187 154 Participants in the UK Biobank. J Crohns Colitis. 2023;17:535–52. doi: 10.1093/ecco-jcc/jjac167. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] 8.Sabino J, Lewis JD, Colombel J-F. Treating Inflammatory Bowel Disease With Diet: A Taste Test. Gastroenterology. 2019;157:295–7. doi: 10.1053/j.gastro.2019.06.027. [DOI] [PubMed] [Google Scholar]

[R9] 9.Halmos EP, Mack A, Gibson PR. Review article: emulsifiers in the food supply and implications for gastrointestinal disease. Aliment Pharmacol Ther. 2019;49:41–50. doi: 10.1111/apt.15045. [DOI] [PubMed] [Google Scholar]

[R10] 10.Marion-Letellier R, Amamou A, Savoye G, et al. Inflammatory Bowel Diseases and Food Additives: To Add Fuel on the Flames! Nutrients. 2019;11:1111. doi: 10.3390/nu11051111. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R11] 11.Martínez Steele E, Juul F, Neri D, et al. Dietary share of ultra-processed foods and metabolic syndrome in the US adult population. Prev Med. 2019;125:40–8. doi: 10.1016/j.ypmed.2019.05.004. [DOI] [PubMed] [Google Scholar]

[R12] 12.Chassaing B, Koren O, Goodrich JK, et al. Dietary emulsifiers impact the mouse gut microbiota promoting colitis and metabolic syndrome. Nature New Biol. 2015;519:92–6. doi: 10.1038/nature14232. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] 13.Chassaing B, Van de Wiele T, De Bodt J, et al. Dietary emulsifiers directly alter human microbiota composition and gene expression ex vivo potentiating intestinal inflammation. Gut. 2017;66:1414–27. doi: 10.1136/gutjnl-2016-313099. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R14] 14.Naimi S, Viennois E, Gewirtz AT, et al. Direct impact of commonly used dietary emulsifiers on human gut microbiota. Microbiome. 2021;9:66. doi: 10.1186/s40168-020-00996-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R15] 15.Chassaing B, Raja SM, Lewis JD, et al. Colonic Microbiota Encroachment Correlates With Dysglycemia in Humans. Cell Mol Gastroenterol Hepatol. 2017;4:205–21. doi: 10.1016/j.jcmgh.2017.04.001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] 16.Winter SE, Thiennimitr P, Nuccio S-P, et al. Contribution of flagellin pattern recognition to intestinal inflammation during Salmonella enterica serotype typhimurium infection. Infect Immun. 2009;77:1904–16. doi: 10.1128/IAI.01341-08. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R17] 17.Chassaing B, Compher C, Bonhomme B, et al. Randomized Controlled-Feeding Study of Dietary Emulsifier Carboxymethylcellulose Reveals Detrimental Impacts on the Gut Microbiota and Metabolome. Gastroenterology. 2022;162:743–56. doi: 10.1053/j.gastro.2021.11.006. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] 18.Daniel N, Wu GD, Walters W, et al. Human Intestinal Microbiome Determines Individualized Inflammatory Response to Dietary Emulsifier Carboxymethylcellulose Consumption. Cell Mol Gastroenterol Hepatol. 2024;17:315–8. doi: 10.1016/j.jcmgh.2023.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] 19.Auchtung JM, Robinson CD, Britton RA. Cultivation of stable, reproducible microbial communities from different fecal donors using minibioreactor arrays (MBRAs) Microbiome. 2015;3:42. doi: 10.1186/s40168-015-0106-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R20] 20.Auchtung JM, Robinson CD, Farrell K, et al. MiniBioReactor Arrays (MBRAs) as a Tool for Studying C. difficile Physiology in the Presence of a Complex Community. Methods Mol Biol. 2016;1476:235–58. doi: 10.1007/978-1-4939-6361-4_18. [DOI] [PubMed] [Google Scholar]

[R21] 21.Schneeberger M, Everard A, Gómez-Valadés AG, et al. Akkermansia muciniphila inversely correlates with the onset of inflammation, altered adipose tissue metabolism and metabolic disorders during obesity in mice. Sci Rep. 2015;5:16643. doi: 10.1038/srep16643. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] 22.Dao MC, Everard A, Aron-Wisnewsky J, et al. Akkermansia muciniphila and improved metabolic health during a dietary intervention in obesity: relationship with gut microbiome richness and ecology. Gut. 2016;65:426–36. doi: 10.1136/gutjnl-2014-308778. [DOI] [PubMed] [Google Scholar]

[R23] 23.Cani PD, Depommier C, Derrien M, et al. Akkermansia muciniphila: paradigm for next-generation beneficial microorganisms. Nat Rev Gastroenterol Hepatol. 2022;19:625–37. doi: 10.1038/s41575-022-00631-9. [DOI] [PubMed] [Google Scholar]

[R24] 24.Martín R, Miquel S, Chain F, et al. Faecalibacterium prausnitzii prevents physiological damages in a chronic low-grade inflammation murine model. BMC Microbiol. 2015;15:67. doi: 10.1186/s12866-015-0400-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R25] 25.Cao Y, Shen J, Ran ZH. Association between Faecalibacterium prausnitzii Reduction and Inflammatory Bowel Disease: A Meta-Analysis and Systematic Review of the Literature. Gastroenterol Res Pract. 2014;2014:872725. doi: 10.1155/2014/872725. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] 26.Miquel S, Martín R, Rossi O, et al. Faecalibacterium prausnitzii and human intestinal health. Curr Opin Microbiol. 2013;16:255–61. doi: 10.1016/j.mib.2013.06.003. [DOI] [PubMed] [Google Scholar]

[R27] 27.Kordahi M, Daniel N, Gewirtz AT, et al. Mucus-Penetrating Microbiota Drive Chronic Low-Grade Intestinal Inflammation and Metabolic Dysregulation. Gut Microbes. doi: 10.1080/19490976.2025.2455790. n.d. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R28] 28.Chassaing B, Ley RE, Gewirtz AT. Intestinal epithelial cell toll-like receptor 5 regulates the intestinal microbiota to prevent low-grade inflammation and metabolic syndrome in mice. Gastroenterology. 2014;147:1363–77. doi: 10.1053/j.gastro.2014.08.033. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R29] 29.Johansson MEV, Gustafsson JK, Holmén-Larsson J, et al. Bacteria penetrate the normally impenetrable inner colon mucus layer in both murine colitis models and patients with ulcerative colitis. Gut. 2014;63:281–91. doi: 10.1136/gutjnl-2012-303207. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R30] 30.Chassaing B, Gewirtz AT. Mice harboring pathobiont-free microbiota do not develop intestinal inflammation that normally results from an innate immune deficiency. PLoS One. 2018;13:e0195310. doi: 10.1371/journal.pone.0195310. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] 31.Mallick H, Rahnavard A, McIver LJ, et al. Multivariable association discovery in population-scale meta-omics studies. PLoS Comput Biol. 2021;17:e1009442. doi: 10.1371/journal.pcbi.1009442. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] 32.Beghini F, McIver LJ, Blanco-Míguez A, et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife . 2021;10:e65088. doi: 10.7554/eLife.65088. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R33] 33.Vieira-Silva S, Falony G, Belda E, et al. Statin therapy is associated with lower prevalence of gut microbiota dysbiosis. Nature New Biol. 2020;581:310–5. doi: 10.1038/s41586-020-2269-x. [DOI] [PubMed] [Google Scholar]

[R34] 34.Levine A, Wine E, Assa A, et al. Crohn’s Disease Exclusion Diet Plus Partial Enteral Nutrition Induces Sustained Remission in a Randomized Controlled Trial. Gastroenterology. 2019;157:440–50. doi: 10.1053/j.gastro.2019.04.021. [DOI] [PubMed] [Google Scholar]

[R35] 35.Sigall Boneh R, Van Limbergen J, Wine E, et al. Dietary Therapies Induce Rapid Response and Remission in Pediatric Patients With Active Crohn’s Disease. Clin Gastroenterol Hepatol. 2021;19:752–9. doi: 10.1016/j.cgh.2020.04.006. [DOI] [PubMed] [Google Scholar]

[R36] 36.Sigall-Boneh R, Pfeffer-Gik T, Segal I, et al. Partial enteral nutrition with a Crohn’s disease exclusion diet is effective for induction of remission in children and young adults with Crohn’s disease. Inflamm Bowel Dis. 2014;20:1353–60. doi: 10.1097/MIB.0000000000000110. [DOI] [PubMed] [Google Scholar]

[R37] 37.Hobson CA, Vigue L, Naimi S, et al. MiniBioReactor Array (MBRA) in vitro gut model: a reliable system to study microbiota-dependent response to antibiotic treatment. JAC Antimicrob Resist . 2022;4:dlac077. doi: 10.1093/jacamr/dlac077. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R38] 38.Hobson CA, Vigué L, Magnan M, et al. A Microbiota-Dependent Response to Anticancer Treatment in an In Vitro Human Microbiota Model: A Pilot Study With Hydroxycarbamide and Daunorubicin. Front Cell Infect Microbiol. 2022;12:886447. doi: 10.3389/fcimb.2022.886447. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R39] 39.Sellem L, Srour B, Javaux G, et al. Food additive emulsifiers and cancer risk: Results from the French prospective NutriNet-Santé cohort. PLoS Med. 2024;21:e1004338. doi: 10.1371/journal.pmed.1004338. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R40] 40.Salame C, Javaux G, Sellem L, et al. Food additive emulsifiers and the risk of type 2 diabetes: analysis of data from the NutriNet-Santé prospective cohort study. Lancet Diabetes Endocrinol. 2024;12:339–49. doi: 10.1016/S2213-8587(24)00086-X. [DOI] [PubMed] [Google Scholar]

[R41] 41.Sellem L, Srour B, Javaux G, et al. Food additive emulsifiers and risk of cardiovascular disease in the NutriNet-Santé cohort: prospective cohort study. BMJ. 2023;382:e076058. doi: 10.1136/bmj-2023-076058. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R42] 42.Iizuka K. Is the Use of Artificial Sweeteners Beneficial for Patients with Diabetes Mellitus? The Advantages and Disadvantages of Artificial Sweeteners. Nutrients. 2022;14:4446. doi: 10.3390/nu14214446. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R43] 43.Markus V. Artificial sweetener-induced dysbiosis and associated molecular signatures. Biochem Biophys Res Commun. 2024;735:150798. doi: 10.1016/j.bbrc.2024.150798. [DOI] [PubMed] [Google Scholar]

[R44] 44.Palmnäs MSA, Cowan TE, Bomhof MR, et al. Low-dose aspartame consumption differentially affects gut microbiota-host metabolic interactions in the diet-induced obese rat. PLoS One. 2014;9:e109841. doi: 10.1371/journal.pone.0109841. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R45] 45.Suez J, Cohen Y, Valdés-Mas R, et al. Personalized microbiome-driven effects of non-nutritive sweeteners on human glucose tolerance. Cell. 2022;185:3307–28. doi: 10.1016/j.cell.2022.07.016. [DOI] [PubMed] [Google Scholar]

[R46] 46.Bonazzi E, Bretin A, Vigué L, et al. Individualized microbiotas dictate the impact of dietary fiber on colitis sensitivity. Microbiome. 2024;12:5. doi: 10.1186/s40168-023-01724-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R47] 47.Caporaso JG, Lauber CL, Walters WA, et al. Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. ISME J. 2012;6:1621–4. doi: 10.1038/ismej.2012.8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R48] 48.Gilbert JA, Meyer F, Jansson J, et al. The Earth Microbiome Project: Meeting report of the “1st EMP meeting on sample selection and acquisition” at Argonne National Laboratory October 6th 2010. Stand Genomic Sci. 2010;3:249–53. doi: 10.4056/aigs.1443528. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] 49.Bolyen E, Rideout JR, Dillon MR, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019;37:852–7. doi: 10.1038/s41587-019-0209-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R50] 50.Callahan BJ, McMurdie PJ, Rosen MJ, et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods. 2016;13:581–3. doi: 10.1038/nmeth.3869. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R51] 51.McDonald D, Price MN, Goodrich J, et al. An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J. 2012;6:610–8. doi: 10.1038/ismej.2011.139. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R52] 52.Jiang L, Huang J, Wang Y, et al. Eliminating the dication-induced intersample chemical-shift variations for NMR-based biofluid metabonomic analysis. Analyst. 2012;137:4209–19. doi: 10.1039/c2an35392j. [DOI] [PubMed] [Google Scholar]

[R53] 53.Tian Y, Zhang L, Wang Y, et al. Age-related topographical metabolic signatures for the rat gastrointestinal contents. J Proteome Res. 2012;11:1397–411. doi: 10.1021/pr2011507. [DOI] [PubMed] [Google Scholar]

[R54] 54.Kechin A, Boyarskikh U, Kel A, et al. cutPrimers: A New Tool for Accurate Cutting of Primers from Reads of Targeted Next Generation Sequencing. J Comput Biol. 2017;24:1138–43. doi: 10.1089/cmb.2017.0096. [DOI] [PubMed] [Google Scholar]

[R55] 55.Inam Z, Felton E, Burrell A, et al. Impact of Antibiotics on the Lung Microbiome and Lung Function in Children With Cystic Fibrosis 1 Year After Hospitalization for an Initial Pulmonary Exacerbation. Open Forum Infect Dis. 2022;9:ofac466. doi: 10.1093/ofid/ofac466. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R56] 56.Hecht G, Bar-Nathan C, Milite G, et al. A simple cage-autonomous method for the maintenance of the barrier status of germ-free mice during experimentation. Lab Anim . 2014;48:292–7. doi: 10.1177/0023677214544728. [DOI] [PubMed] [Google Scholar]

[R57] 57.Katakura K, Lee J, Rachmilewitz D, et al. Toll-like receptor 9-induced type I IFN protects mice from experimental colitis. J Clin Invest. 2005;115:695–702. doi: 10.1172/JCI22996. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R58] 58.Chassaing B, Srinivasan G, Delgado MA, et al. Fecal lipocalin 2, a sensitive and broadly dynamic non-invasive biomarker for intestinal inflammation. PLoS One. 2012;7:e44328. doi: 10.1371/journal.pone.0044328. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R59] 59.Johansson MEV, Hansson GC. Preservation of mucus in histological sections, immunostaining of mucins in fixed tissue, and localization of bacteria with FISH. Methods Mol Biol. 2012;842:229–35. doi: 10.1007/978-1-61779-513-8_13. [DOI] [PubMed] [Google Scholar]

PERMALINK

Invitro microbiota model recapitulates and predicts individualised sensitivity to dietary emulsifier

Héloïse Rytter

Sabrine Naimi

Gary Wu

Jim Lewis

Maeva Duquesnoy

Lucile Vigué

Olivier Tenaillon

Eugeni Belda

Marta Vazquez-Gomez

Nina Touly

Djésia Arnone

Fuhua Hao

Ruth E Ley

Karine Clément

Laurent Peyrin-Biroulet

Andrew D Patterson

Andrew T Gewirtz

Benoit Chassaing

Abstract

Background

Objectives

Design

Results

Conclusion

WHAT IS ALREADY KNOWN ON THIS TOPIC

WHAT THIS STUDY ADDS

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

Introduction

Results

MBRA model faithfully recapitulated microbiotas of the FRESH subjects

MBRA reproduced the individualised CMC-sensitivity status observed in the FRESH study

Prediction of CMC sensitivity status via metagenomic signature and MBRA model

In vivo validation of CMC-sensitivity predictions

CMC sensitivity associated with a baseline microbiota signature within the whole FRESH cohort

Prevalence of metagenomic markers of emulsifier sensitivity in independent cohorts of healthy participants and patients

Discussion

Materials and methods

Reagents

Faecal sample collection

Setup and use of the in vitro MBRA system and treatment with CMC

Bacterial DNA extraction

Bacterial density quantification through 16S rRNA qPCR

Microbiota analysis by 16S rRNA gene sequencing using Illumina technology

16S rRNA gene sequences analysis

1H NMR-based metabolomic analysis

Microbiota analysis by shotgun metagenomic using Illumina technology

Mice experiments

Histopathological analysis of colonic tissue and immunochemistry

Immunostaining of mucins and localisation of bacteria by fluorescent in situ hybridisation

Statistical analysis

Patient and public involvement

Supplementary material

Acknowledgements

Footnotes

Data availability statement

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

¹H NMR-based metabolomic analysis