Dietary fibre is good for you, except when it’s not


Heather K. Armstrong, Michael Bording-Jorgensen, Deanna M. Santer, Zhengxiao Zhang, Rosica Valcheva, Aja M. Rieger, Justin Sung-Ho Kim, Stephanie I. Dijk, Ramsha Mahmood, Olamide Ogungbola, Juan Jovel, France Moreau, Hayley Gorman, Robyn Dickner, Jeremy Jerasi, Inderdeep K. Mander, Dawson Lafleur, Christopher Cheng, Alexandra Petrova, Terri-Lyn Jeanson, Andrew Mason, Consolato M. Sergi, Arie Levine, Kris Chadee, David Armstrong, Sarah Rauscher, Charles N. Bernstein, Matthew W. Carroll, Hien Q. Huynh, Jens Walter, Karen L. Madsen, Levinus A. Dieleman, Eytan Wine

Research paper: Unfermented β-fructan fibers fuel inflammation in select inflammatory bowel disease patients (

Media release: Dietary fibre is good for you, except when it’s not | Folio (


Digestible carbohydrates are degraded in the small intestine; non-digestible carbohydrates (fiber and resistant starch) are fermented by colonic microbes. Fermentation of dietary fibers produces gases, lactate, and short chain fatty acids (SCFAs), with multiple beneficial physiological effects, but fibers have also been shown to be harmful in select situations. The beneficial potential for fermentable fibers in inflammatory bowel diseases (IBD) is demonstrated by low SCFA production, especially in ulcerative colitis (UC), linked to absence of SCFA-producing microbes. Administration of β-fructan fibers improved mild UC, associated with increased SCFA (butyrate) production. However, this generally positive effect of fibers related to fermentation and SCFA production seems to have overshadowed potential detriments, as many patients with IBD describe sensitivity to fiber consumption; ignoring or not understanding this process can lead to avoiding non-digestible fibers altogether through exclusion diets. Such exclusion diets can improve symptoms but may deprive patients of the benefits of fibers, which are especially important in IBD.

While research has intensified on the role of fiber fermentation in IBD, and β-fructan fibers in particular have been gaining attention for their prebiotic potential (promoting growth of ‘beneficial microbes’), the role of microbiota and fiber fermentation processes, and whether they are beneficial or detrimental, remains poorly understood. Structurally, dietary fibers (Table S1) and cell wall components of microorganisms (e.g., fungal β- (1,3)glucans) are polymers of greater than 3 sugars (oligofructose [FOS] ~8 sugars; grain β-D-glucan ~3 sugars) and ranging up to 50-100 sugars (inulin; fungal β-(1,3)glucan), which can vary in their degree of polymerization (DP), branching, solubility, and interactions with host cells. Immune response to polysaccharides on the surface of fungal cells suggests a possible link between whole unfermented fibers and inflammation. β-(1,3) glucan on the surface of fungi (e.g., zymosan, curdlan) interacts with immune cells (e.g., macrophages), inducing pro-inflammatory antifungal immunity via Dectin-1 and TLR2. Similarly, β-fructan fibers (inulin and FOS) induce TLR-mediated inflammatory pathways. This led us to hypothesize that in patients with reduced fiber-fermenting microbes (e.g., IBD), dietary fibers could remain intact, interact with host cell receptors, and promote gut inflammation. Here we demonstrate that unfermented dietary β-fructans induce pro-inflammatory cytokine secretion in select IBD patients, mediated by microbial functions. Our data suggest that select fibers may be detrimental in individuals lacking fermentative microbes (e.g., IBD, other chronic illnesses, antibiotic use), with increased opportunity for interactions between host immune cells and luminal contents (due to increased immune cells and disrupted epithelial barrier). These same fibers provide health benefits in individuals with high fermentative potential.


IBD patients describe variable intolerance of fiber consumption, which can lead to avoidance of generally beneficial fibers and worse patient outcomes. We utilised IBD as a model to confirm our hypothesis that fibers that remain unfermented could drive inflammation. Supporting our findings, β-fructans have been shown to induce reactive oxygen species (ROS) production and associated inflammation, possibly through NLRP3 signalling and clinical studies show that FOS consumption can worsen outcomes in select patients. Nevertheless, the potentially negative effects of dietary fibers are poorly documented and usually overlooked. Here, unfermented FOS induced pro-inflammatory cytokines in PBMCs, THP-1 macrophages, and IBD patient biopsies cultured ex vivo, via pathways previously associated with IBD including TLR2 and NLRP3, which are known to corporate in response to ligand stimuli such as lipopolysaccharide (LPS). This was confirmed in select IBD patients in an RCT inulin/FOS-treated cohort. We propose that interactions between unfermented fibers in the luminal contents with leukocytes found in the mucosal lining or lamina propria exposed due to epithelial barrier breakdown, could drive these responses in a physiological setting.

TLR2 is differentially expressed in cell types, with greater levels in monocytes and macrophages, supporting increased potential for interaction between unfermented β-fructans and TLR2 in IBD patients where these cell populations are increased, particularly in patients with active disease. Response could be explained further by the increased presence of inflammatory macrophages in inflamed tissues of IBD patients although the presence of specific macrophage populations was not examined in this study. The epithelial barrier, which typically prevents undesired immune interaction with luminal content, is commonly disrupted in IBD. Here, β-fructans improved barrier formation in vitro, while β-D-glucan reduced barrier formation, possibly due to structural differences between β-fructans and β-D-glucan. Our data support future investigation of these pathways using organoid and animal models which would provide more mechanistic findings in relation to the effects of these fibers on epithelial barrier integrity. Sensitivity to fibers was further supported by our findings that patients with active disease (IBD-R) consumed less dietary fiber than patients in remission (IBD-NR), and lower FOS consumption (measured by FFQ) correlated with higher pro-inflammatory response to FOS in matching patient biopsies cultured ex vivo. This suggests that patients with fiber sensitivities might unknowingly avoid consumption of select FOS-containing foods, possibly in attempts to ameliorate symptoms.

Microbial function was predictive of response to FOS in patient samples. The enzyme UDP-N acetylglucosamine was increased in IBD-R, suggesting that this pathway may be involved in the pro-inflammatory response to fibers, possibly via T-cell activation. In contrast, IBD-R had significantly reduced riboflavin synthase which displays anti-inflammatory, antioxidant, and microbe-altering properties in IBD patients. Both riboflavin synthase and glucosylceramidase inhibit a variety of pro-inflammatory cytokines (IL-6, TNF, IL-1β). Riboflavin (vitamin B2) was lower in stool of UC patients in our RCT who flared following 6-month consumption of β-fructans and its absence correlated with increased fold-change in fecal calprotectin, suggesting further links between riboflavin and response to fiber consumption in IBD patients. Faecalibacterium prausnitzii is thought to use riboflavin as a mediator of butyrate and SCFA production, suggesting a key link between fiber-fermenting microbes, enzyme abundance, SCFA production, and inflammatory response to dietary fibers. IBD-R mucosal microbiota washes produced increased acetate and decreased propionate and butyrate through fermentation of FOS. Acetate is known to increase ROS production in macrophages, while butyrate and propionate inhibit inflammation through various pathways, suggesting that even when fermentation is not reduced, altered production of SCFA may promote inflammation.

While no individual microbe species associated explicitly with pro-inflammatory response to FOS, there were altered patterns of microbial species abundances that may help identify microbiome changes associated with altered fermentation. Abundance of microbes known to ferment fibers was significantly reduced in moderate and severe CD patients; particularly, the dominant fiber fermenting and butyrate-producing microbes Roseburia hominis and F. prausnitzii, as expected. While recent studies have indicated the importance of these mucosal microorganisms in the gut ecosystem and in relation to diet, there is only limited research on the interactions of diet with mucosal microbiota in IBD. It is important to note some limitations of our study. While we confirmed the purity of the fibers used with low LPS (within test limits of detection), other microbial contaminants may co-purify with the three β-fructans (FOS/inulin) from chicory roots used in this study, along with other fibers. Further, samples were collected following colonoscopy preparation, which is known to alter microbiota composition; while our main focus was on mucosa associated microbes, which are less impacted by bowel prep, luminal microbiota also play an important role in fibre fermentation.

Mucosal microbes typically include important fiber fermenting microbes (Bacteroidetes, Firmicutes [Veillonellaceae, Ruminococcaceae) compared to stool microbiota; however, our understanding of the precise community of microbes (luminal or mucosal) involved in fiber fermentation remains limited by our ability to culture and identify these microbes. We propose that when fiber-fermenting microbes are present in the gut, and normal barrier integrity prevents interactions between fibers and underlying immune cells, fermentation of select fibers enhances the barrier and reduces inflammatory response. In contrast, select disease state scenarios, such as active IBD, provide conditions leading to increased exposure and sensitivity to unfermented dietary fibers to develop in the diseased gut microenvironment. These conditions include 1) a reduced abundance and capacity of the gut microbiota to ferment fiber, 2) increased presence of immune cells at the mucosal surface, and 3) inflammatory damage to the gut barrier. Interaction of FOS and inulin with host cells could then result in gut inflammation through direct effects of intact fiber and/or altered SCFA production. Our work could have significant impacts on patient care. Further clinical studies are warranted to determine if FOS should be avoided by IBD patients when experiencing specific alterations in gut microbiota composition and functions, specifically associated with lack of fermentation, especially with active disease. Since altered microbiota is more frequently found in active IBD patients, it could be speculated that FOS (and potentially other fibers) should be administered as adjunct therapy only after medical therapy has induced remission (with barrier repair/mucosal healing and healthy microbiota functions) in these individuals, to ensure the other benefits of fibers and their products.