Edited by: Clara G. De Los Reyes-Gavilan, Consejo Superior de Investigaciones Científicas (CSIC), Spain
Reviewed by: Raylene A. Reimer, University of Calgary, Canada; Ravinder K. Nagpal, Florida State University, United States
This article was submitted to Nutrition and Microbes, a section of the journal Frontiers in Nutrition
†These authors have contributed equally to this work and share the first authorship
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Human gut microbiota has a fundamental role in human health, and diet is one of the most relevant factors modulating the gut microbial ecosystem. Fiber, fat, proteins, and micronutrients can shape microbial activity and structure. Much information is available on the role of defined prebiotic fibers on gut microbiota, but less known are the effects of intact dietary fiber sources on healthy gut ecosystems. This research investigated
Dietary fiber is composed of carbohydrate polymers with three or more monomeric units, which are not digested or absorbed in the human small intestine (
Dietary fiber intake has been associated with multiple health benefits, including reduced risk for heart disease, stroke, hypertension, specific gastrointestinal disorders, obesity, type 2 diabetes, and some cancers (
Due to health benefits associated with dietary fiber intake, multiple organizations such as the UK Scientific Advisory Committee on Nutrition (
Reasons for the fiber intake shortfall are multifactorial, including consumer misperception on dietary sources of fiber and adequate intake amounts, changing dietary habits toward Western, high-protein, gluten-free, wheat-free, or grain-free diets, or gastrointestinal discomfort associated with gas production after increasing fiber intake (
Some of the health benefits of dietary fiber can be linked to the gut microbiota. Dietary fibers undergo bacterial fermentation upon arriving in the colon and thus impact the composition and functionality of bacterial communities, including the production of fermentative metabolites with different effects on the host (
Whole grains, vegetables, legumes, and fruit intrinsically contain fiber and provide most of the fiber to the diet. The major cereals consumed worldwide are wheat, milled rice, and maize, followed in production by oat, rye, barley, triticale, millet, and sorghum. Cereals are the leading staple food and considering that whole grain cereals contain an average amount of 7% of TDF could provide the recommended daily intake of fiber per day if consumed in the appropriate form. Other fiber source groups like pulses have declined in consumption per capita due to changes in dietary changes, consumer preferences, and domestic production failure in developing countries (
Additionally, isolated and synthesized fibers can be added to foods or consumed as food supplements, to augment intake, with beneficial physiological effects. Prebiotics are a specific type of fiber, defined as a substrate that is selectively utilized by host microorganisms conferring a health benefit (
We selected a wide variety of food sources of fiber (22 products) and tested in parallel the effect on gut microbiota activity [pH, gas, SCFA, lactate] and specific composition of Firmicutes, Bacteroidetes, bifidobacteria, and lactobacilli
Description of the 22 fiber sources used in gastrointestinal digestion and colonic fermentation is shown in
Description of products used in this research.
Whole grain cereals (WG) | WG1 | Whole grain millet | Woodland Foods | G07 | Whole grain | 15 lb batch of grains cooked with water and steam in a Kellogg-designed pressurized rotary cooker for 50 min at 121°C. Then dried on a Kellogg-designed bed dryer with forced air at 116°C for 90 min. |
WG2 | Whole grain oat | Richardson Milling | RM-G001 | Whole grain | ||
WG3 | Medium grain whole brown rice | Gulf Rice | BRMG2-4 | Whole grain | ||
WG4 | Non waxy whole grain soft white wheat | Adams | G1290 | Whole grain | ||
WG5 | Whole grain barley (tamalpais/hulless) | Adams | G1444 | Whole grain | ||
WG6 | Whole grain corn | Grain Millers | F75WG | Whole grain | ||
WG7 | Waxy whole grain soft white wheat | Adams | G1448 | Whole grain | ||
WG8 | Waxy hulless barley | Adams | 4000 | Whole grain | ||
Cereals (non-whole grain) (C) | C1 | Oat beta glucan | Lantmännen | PromOat | Oat Bran, Nordic Oats | Heated at 121°C using a Lincoln Impinger oven for 10 min. |
C2 | Rice fiber | Rettenmaier | Vitacel RF310 | Hull | ||
Seeds (S) | S1 | Whole brown flaxseed | Hesco /HFI | CS362 | Whole seed | Dried on a Kellogg-designed bed dryer with forced air at 116°C for 30 min. |
S2 | Hemp seed | Woodland Foods | N73 | Shelled hemp seed | ||
S3 | Psyllium fiber | Perrigo | 125PH-1-1 | Seed husk | Heated at 121°C using a Lincoln Impinger oven for 10 min. | |
Pulses (P) | P1 | Whole brown lentils | Woodland Foods | B27 | Whole lentil | 15 lb batch of grains cooked with water and steam in a Kellogg-designed pressurized rotary cooker for 50 min at 121*C. Then dried on a Kellogg-designed bed dryer with forced air at 116*C for 90 min. |
P2 | Soy fiber | Dupont | Fibrim 2000 | Cell wall material of soybean cotyledon | Heated at 121°C using a Lincoln Impinger oven for 10 min. | |
P3 | Pea fiber | Roquette | I 50 M | Hull of yellow pea | ||
Other fibers (F) | F1 | Kiwi fiber | AIDP | Livaux 5228-00-F | Gold Kiwi Fruit. Peeled, seeds removed. Freeze-dried flesh | Not processed at all—heat-sensitive |
F2 | Inulin | Beneo | Orafti GR | Chicory roo | Heated at 121°C using a Lincoln Impinger oven for 10 min. | |
F3 | Bamboo fiber | Natural Fiber Solutions | Unicell BF200 | Whole plant crushed, fiber-extracted | ||
F4 | Konjac flour | Dupont | Nutricol ME 8731 | Root of tuber | ||
F5 | Apple fiber | Natural Fiber Solutions | AP200 | Crushed apples, fiber-extracted. | ||
F6 | Algal beta glucan isolate | Noblegen | Eunite |
For each product, nutritional values reported in the product datasheets were recorded and summarized in
Fifteen of the 22 test products (
The
After the gastric incubation, small intestinal fluid was added to the simulated gastric chyme (1:1 v/v), including pancreatin from porcine pancreas (100 U ml−1 of trypsin in the final mixture) (Sigma) and bile salts (10 mM, Oxoid), pH adjusted to 5.5 and incubated at 90 rpm for 30 min, simulating the duodenal digestion. Then, a dialysis approach was applied to simulate the small intestinal absorption. Briefly, 3.5-kDa dialysis membrane (ZelluTrans/Roth dialysis membranes, Carl Roth) was filled with small intestinal digest adjusted to pH 7 and submerged in dialysis fluid (3.75 g/L NaHCO3, pH 7) at 37°C on a shaker for 3 h, with the adjustment of pH of the intestinal content and replacement of the dialysis fluid every 45 min. At the end of the dialysis, the remaining intestinal solution was weighed and used for subsequent colonic incubations. Digestions were performed in single reactors for each product.
One of the products (F1, kiwi fiber), with a sugar content above 10 g/100 g, was dialyzed through a 0.5-kDa membrane without gastrointestinal predigestion.
The short-term screening assay consisted of a simulated colonic incubation of a single dose of the selected fiber source under conditions representative for the proximal colon region of an adult human, using representative bacterial inocula from three healthy donors (
Overview of the gastrointestinal digestion and colonic incubation
Short-term fecal batch incubations were performed as previously described in Van den Abbeele et al. (
Samples collected after 24 and 48 h of incubation were evaluated for the total amount of Firmicutes, Bacteroidetes, bifidobacteria, and lactobacilli by qPCR. The DNA was extracted from a pellet of bacterial cells originated from a 1 ml sample after centrifugation for 5 min at 7,700 g. A Fastprep-24 device (MP BioMedicals, Illkirch, France) was used for homogenization (two cycles of 40 s at 4 m/s). Subsequently, the qPCR assays were performed using a StepOnePlus Real-Time PCR system (Applied Biosystems, Foster City, CA), using the primers and conditions described in
The pH (Senseline F410; ProSense, Oosterhout, The Netherlands), gas (hand-held pressure indicator CPH6200; Wika, Echt, The Netherlands), lactate, and short-chain fatty acid (SCFA) measurements were taken at 0, 6, 24, and 48 h after starting the colonic incubation. Acetate, propionate, butyrate, and branched SCFAs (isobutyrate, isovalerate, and isocaproate) were measured as described by De Weirdt et al. (
Each donor (n = 3)/product (n = 22) combination was tested in single reactors, and different donors were considered biological triplicates. To calculate statistically significant differences between groups of products by category, data from each category (control, whole grain cereals, non-whole grain cereals, seeds, pulses, other fibers) were pooled.
Delta values for pH, gas, SCFA, BCFA, and copies/ml obtained from qPCR were calculated by subtracting values obtained at the end of the experiment (48 h) minus values at the starting of the experiment (0 h) (Δ48 h). Relative values to control condition were calculated by the values obtained in each treatment minus the values in the control condition (rΔ48 h). Firmicutes/Bacteroidetes ratio was calculated using Δ48-h data.
All statistical analyses were performed in GraphPad Prism 9.0.0 (GraphPad Software, San Diego, CA) in relative Δ48 h. Significance level was set at 0.05. Normality of the dataset was tested with the Kolmogorov–Smirnov test. In the case of normality, the mean values of two different groups were compared with an independent samples t-test. Significant differences in microbial metabolites between different groups of fibers were tested with one-way ANOVA in the case of normality. Homogeneity of variances was tested with the modified Levene's test. Depending on the outcome of the Levene's test, Bonferroni or Dunnett's T3 was used as
Principal component analysis (PCA) plots were obtained by centering and scaling dimensions in ClustVis (
The whole grain cereal group showed a homogenous nutritional composition, with values of protein (10.9 ± 2.4%), fat (3.2 ± 1.6%), dietary fiber (12.5 ± 5.7%), sugars (0.6 ± 0.3%), and other carbohydrates (60.9 ± 8.1%) in the same range for all the different products (
The other fiber groups were compositionally heterogeneous. Within the non-whole grain cereals, C1 (oat beta glucan) had a 33% lower fiber content (65.0%) than C2 (rice fiber) (93.1%). Within the pulse group, P2 (soy fiber) and P3 (pea fiber) had a similar nutritional content in terms of protein (8–12%), fat (1–1.1%), dietary fiber (75–82%), sugars (n.d.−0.1%), and other carbohydrates (n.d.−3%). P1 (whole brown lentils), however, had a higher content of protein (24.6%) and other carbohydrates (50.6%). In the seeds group, the highest fat content (45.5 ± 4.7%) was found in S1 (whole brown flaxseed) and S2 (hemp seed). S3 (psyllium) showed a different nutritional profile compared to other seeds, with lower fat (0.8%) and protein (2.7%), and higher fiber content (97.1%).
When comparing dietary fiber levels described in the product datasheets with total fiber content quantified by the McCleary method, results were typically in agreement, with variation <20% (
Considering the PCA plot including metabolic (gas, pH, SCFA, BCFA, lactate) and qPCR data (Firmicutes, Bacteroidetes, bifidobacteria, lactobacilli), control condition without fermentable substrates is distant from most of the products, except C2 (rice fiber), F3 (bamboo fiber), and F6 (algal beta glucan isolate) (
Principal component analysis (PCA) plots representing metabolic (pH, gas, SCFA, BCFA, lactate) and qPCR (bifidobacteria, lactobacilli, Bacteroidetes, Firmicutes) data obtained for batch assays with fecal samples of three healthy donors exposed to control and 22 fiber products. BCFA, branched-chain fatty acid; SCFA, short-chain fatty acid. WG1, whole grain millet; WG2, whole grain oat; WG3, medium grain whole brown rice; WG4, whole grain soft white wheat; WG5, whole grain barley; WG6, whole grain corn; WG7, waxy whole grain soft white wheat; WG8, waxy hulless barley; C1, oat beta glucan; C2, rice fiber; S1, whole brown flaxseed; S2, hemp/hemp hearts; S3, psyllium fiber; P1, whole brown lentils; P2, soy fiber; P3, pea fiber; F1, kiwi fiber; F2, inulin; F3, bamboo fiber; F4, konjac flour; F5, apple fiber; F6, algal beta glucan isolate.
S1 (whole brown flaxseed) and S2 (hemp seed) are both high-fat-containing products, with fat% >10 and >40%, respectively, and had an intermediate pattern between the control and whole grains. S3 (psyllium fiber) had a different effect, not comparable to other groups or individual fibers, mainly due to a strong stimulation of propionate production, likely due to fermentation by increased levels of Bacteroidetes.
Within the entire group of test products, there was a high intragroup variability. F2 (inulin) is distant from other groups and control condition, indicating it is strongly fermentable, inducing significant lactate accumulation, BCFA inhibition, propionate production, and stimulatory potential on bifidobacteria and lactobacilli. F1 (kiwi fiber), F4 (konjac flour), and F5 (apple fiber) clustered more centrally within these extremes.
Overall, fiber supplementation promoted bifidobacteria (
Effect of different groups of dietary fibers on specific members of gut microbiota.
Bifidobacteria | 0.29 ± 0.25 | 0.11 ± 0.14 | 0.33 ± 0.34 | ||
Lactobacilli | 0.11 ± 0.24 | 0.09 ± 0.39 | 0.62 ± 1.00 | ||
Firmicutes | 0.35 ± 0.29 | ||||
Bacteroidetes | 0.17 ± 0.12 | 0.45 ± 0.34 | 0.12 ± 0.25 | ||
Ratio F/B | 1.00 ± 0.33 | 0.96 ± 0.03 | 0.98 ± 0.02 | 1.00 ± 0.03 |
Effect of dietary fibers on absolute abundance of bifidobacteria, lactobacilli, Bacteroidetes, and Firmicutes. Bars represent the increase of bifidobacteria
Within the non-whole grain cereal group, only C1 (oat beta glucan) increased the levels of bifidobacteria, lactobacilli, and Firmicutes and clustered close to whole grains (
When analyzed together, seeds did not affect the levels of bifidobacteria, lactobacilli, or Bacteroidetes (
Within the pulse group, P1 (whole brown lentils) and P3 (pea fiber) induced significant increases in bifidobacteria, lactobacilli, Bacteroidetes, and Firmicutes, clustering closer between them than to P2 (soy fiber), which only influenced Bacteroidetes group (
Within the other fiber group, there was a larger intergroup variation. F1 (kiwi fiber) and F2 (inulin) increased bifidobacteria, lactobacilli, and Firmicutes. F1 exerted marked bifidogenic effects and clustered close to the whole grain products. Remarkably, F2 is differentiated from all the other samples in the PCA plot, likely due to the highest effect on lactobacilli populations. F3 (bamboo fiber) and F6 (algal beta glucan isolate) had milder effects on microbial composition and grouped closer to the control condition than other fibers. F4 (konjac flour) and F5 (apple fiber) clustered in between due to intermediate effects, with F5 stimulating bifidobacteria, Firmicutes, and Bacteroidetes and F4 only increasing Bacteroidetes and Firmicutes members. F6 did not affect any of the groups, and it is the closest sample to the control (
WG3 (medium grain whole brown rice), WG8 (waxy hulless barley), P1 (whole brown lentils), and F2 (inulin) significantly increased the Firmicutes/Bacteroidetes ratio compared to control (
When considering the different products in fiber groups, whole grains and pulses induced the highest pH drop and gas production, with significant changes of pH (−0.5 ± 0.1,
Effect of different groups of dietary fibers on bacterial metabolic markers.
pH | −0.2, 0.0 | −0.4, 0.2 | −0.3, 0.2 | – |
−0.5, 0.3 | |
Gas (kPa) | 28.2, 1.1 | 42.3, 13.4 | ||||
Lactate (mM) | 1.2, 0.1 | 5.5, 5.6 | 1.8, 0.7 | 3.5, 2.4 | 4.5, 6.9 | |
Acetate (mM) | 22.4, 4.2 | 33.9, 10.8 | ||||
Propionate (mM) | 8.0, 2.1 | 16.0, 3.6 | 13.7, 6.3 | 12.2, 3.8 | ||
Butyrate (mM) | 3.1, 2.1 | 7.3, 4.6 | 6.0, 5.7 | 4.3, 1.7 | 5.6, 3.3 | 4.5, 2.4 |
Branched FA (mM) | 1.9, 1.2 | 1.6, 1.4 | 1.8, 1.3 | 2.2, 1.6 | 1.7, 1.3 | 1.6, 1.2 |
Total SCFA (mM) | 36.5, 9.1 | 70.6, 5.8 | 56.2, 17.9 | 56.3, 11.8 | 66.4, 17.8 | 53.5, 13.1 |
The highest gas production and pH decrease were observed for F2 (inulin), while the lowest effect on pH and gas production was observed for S2 (hemp, −0.16 ± 0.01) and C2 (rice fiber, 31.4 ± 3.9 kPa), respectively.
In general, fibers induced a significant increase (ANOVA,
Total SCFA were consistently increased in all the groups (53.5–70.6 mM, ANOVA
Effect of dietary fibers on microbial activity. Bars represent the net values of pH
Regarding butyrate, it was a trend to increase levels by the supplementation of fibers (5.8 ± 3.7 mM) compared to control (3.1 ± 2.1 mM); however, the changes were not always significant due to a high interindividual variability on butyrate production at basal levels. In fact, the microbiota from one of the donors included in the study had low butyrogenic potential, with levels of butyrate of 1.9 ± 0.8 mM (average for all the treatments), compared with the other donors (5.9–8.8 mM) (
This study assessed the effect of 22 commercially available food sources of fiber on gut microbial composition and function of three healthy human donors
We tested a wide variety of fiber sources, including whole grain cereals, cereals (non-whole grain), seeds, pulses, and other fibers; however, a high heterogenicity in terms of composition was observed within each group. The groups with nutrients available for absorption in the small intestine were predigested, with intestinal absorption mimicked by a dialysis approach, to ensure that non-digestible compounds, and not the source of easily available sugars derived from product digestion, were the driver of microbial shifts in the colonic fermentation.
The group of whole grain products had a consistent effect in increasing acetate, lactate, and propionate after short-term incubations, together with the stimulation of bifidobacteria and lactobacilli. Firmicutes members were fueled by whole grain products, while Bacteroidetes group was, for most of the whole grain fibers, not significantly affected. These results are in agreement with human intervention studies showing that whole grains can impact fecal microbe levels; e.g., wheat, oat, and corn can increase
Lactate was one of the metabolites specifically and consistently induced by whole grains, together with P1 (whole brown lentils) and F2 (inulin). Lactate is produced by many microbial members of the human gut, including lactic acid bacteria and bifidobacteria. lactate can influence colonic pH and has been shown to inhibit the growth of some pathogenic bacteria, including
The effects on fermentation profiles and gut microbiota modulation of oat β-glucan were similar to those observed in whole grain cereals, despite the amount of dietary fiber in oat beta glucan being approximately five times higher. Oat β-glucans are classified as prebiotic soluble fibers with recognized health benefits (
The products with the lowest fermentability were rice fiber, bamboo fiber, and algal beta glucan isolate. These products were high in fiber, rich in either soluble (algal beta glucan) or insoluble fiber, with a limited amount (<1%) of other macronutrients such as protein, fat, sugars, or other non-fiber/sugar carbohydrates. This group seems to be undigested by the enzymatic repertoire in the human gut microbiota; however, a larger screening including more donors would be required for consistent conclusions. Although largely unfermented, this does not imply that they lack potential physiological benefit. Recently, bamboo shoot fiber has shown the ability to prevent obesity in high-fat-fed mice through the modulation of the gut microbiota (
The interindividual variability on taxonomic and functional composition of the gut microbiota is a key element in responses to dietary interventions (
We analyzed three different microbial ecosystems derived from healthy donors in response to dietary fiber addition and obtained similar trends for all the donors. We identified the limitation of using a low number of donors and proposed larger screening with the most promising dietary fibers, including more donors to confirm our results. Further studies, including multiple fibers and human-derived disease-dysbiotic microbial communities, like occurring in obesity, diabetes, or inflammatory bowel disease, would provide more evidence on fiber diversity on microbial modulation of dysbiotic gut ecosystems.
Remarkably, bifidobacteria populations were stimulated in all the donors by all whole grains, brown lentils, pea fiber, kiwi, and apple fiber. The potential to promote specific individual bifidobacteria within individual gut ecosystems with diet-based solutions may benefit gut homeostasis “from the inside,” targeting residential strains already co-evolved with the host (
Previous research using mice models has shown that fiber deprivation triggers the colonic microbiota to utilize host-secreted glycoproteins as a nutrient source (
In addition to the relevant effect of fiber sources on mucosal health, complex fiber structures such as wheat bran can serve as a microenvironment to be colonized by specific degraders (
The authors acknowledge significant limitations associated with this research, including (i) limited number of donors; (ii) short-term evaluation of fiber effect in static conditions; (iii) the microbiota analysis based on qPCR data of specific members of the gut microbiota offering less resolution than complete 16S rRNA gene sequencing; and (iv) lack of full characterization of nutritional parameters such as specific fiber structure or micronutrient contents.
Despite these limitations, our
The original contributions presented in the study are included in the article/
The studies involving human participants were reviewed and approved by University Hospital Ghent (reference number B670201836585). The patients/participants provided their written informed consent to participate in this study.
PV and AB were involved in conceptualization. MC, PV, JG, and AB were involved in data curation. PV, MC, and JG were involved in formal analysis. PV and AB were involved in funding acquisition. PV and JG were involved in methodology. PV and MM were involved in project administration. ER and AB were involved in utilizing resources. MM was involved in supervision. MC and PV were involved in visualization. MC was involved in writing the original draft. MC and AB were involved in writing the review and editing. All authors contributed to the article and approved the submitted version.
AB and ER are employees of the Kellogg Company. PV was employed by ProDigest. MC and JG are employed by ProDigest. MM is CEO of ProDigest. The authors declare that this study received funding from Kellogg Company. The funder had the following involvement in the study: conceptualization, funding acquisition, data curation, utilizing resources, writing and editing.
The Supplementary Material for this article can be found online at: