ArticlePDF Available

A low-gluten diet induces changes in the intestinal microbiome of healthy Danish adults

Authors:
Article

A low-gluten diet induces changes in the intestinal microbiome of healthy Danish adults

Abstract and Figures

Adherence to a low-gluten diet has become increasingly common in parts of the general population. However, the effects of reducing gluten-rich food items including wheat, barley and rye cereals in healthy adults are unclear. Here, we undertook a randomised, controlled, cross-over trial involving 60 middle-aged Danish adults without known disorders with two 8-week interventions comparing a low-gluten diet (2 g gluten per day) and a high-gluten diet (18 g gluten per day), separated by a washout period of at least six weeks with habitual diet (12 g gluten per day). We find that, in comparison with a high-gluten diet, a low-gluten diet induces moderate changes in the intestinal microbiome, reduces fasting and postprandial hydrogen exhalation, and leads to improvements in self-reported bloating. These observations suggest that most of the effects of a low-gluten diet in non-coeliac adults may be driven by qualitative changes in dietary fibres.
A low-gluten diet alters the composition of the gut microbiome. a Scatterplot of the statistical significance of the metagenomic species (MGSs) as assessed by a linear mixed model testing for the difference between the low-gluten and the high-gluten diets adjusted for age, gender, intestinal transit time, participant (n = 51) and carry-over effect. Adjusted P values are displayed on the y-axis (log10 scale) and the effect size (absolute values were log10 transformed) is on the x-axis. Points are sized according to the total abundance (%) and coloured according to the ten most abundant taxonomic families. The ‘Other’ category consists of the remaining families. The horizontal line represents an adjusted P value of 0.05 and the 14 species that changed significantly (FDR < 0.05) between the interventions are labelled with their full taxonomic annotation. Only species that could be annotated to family level and with abundance above 0.02% were included in the plot (255 species). b Bar chart of the 14 significant species showing the log2 fold change (means ± SEM) between baseline and after the low-gluten diet (blue bars) and high-gluten diet (red bars), respectively. The black circles are sized according to the negative log10 of the adjusted P values of comparison between the low-gluten and the high-gluten diet using a linear mixed model adjusted for age, gender, intestinal transit time, participant (n = 51) and carry-over effect. Green circles are scaled according the species abundance. The last column lists the number of participants in whom the given species were measured. Details on the individual species can be found in Supplementary Data 2
… 
A low-gluten diet affects measures of intestinal fermentation. a Breath hydrogen levels following the same standardised meal at all four visits (low-gluten diet start, open blue circles; low-gluten diet end, blue squares; high-gluten diet start, open red triangle; high-gluten diet end, filled red triangle). Data are shown as means ± SEM (n = 51-57). b Plot showing changes in gut bloating as assessed by visual analogue scale (VAS) following the low-gluten diet (blue circles) compared with the high-gluten diet (red squares). Data are shown as means ± SEM (n = 52–53). Changes were assessed by a linear mixed model adjusting for age, gender and intestinal transit time. *P < 0.05, **P < 0.01. c Linear regression network of breath hydrogen levels and the abundance of bacterial species and concentrations of urine metabolites which are significantly responding to the dietary interventions using a linear mixed model adjusted for gender, age and participant (n = 49) (Supplementary Data 5). The dotted line separates the features that were decreased and increased, respectively, when comparing the low-gluten and high-gluten periods. Significant (FDR < 0.05) positive associations are indicated with grey lines; negative associations with red lines. Thickness of lines indicates the significance level. Nodes are coloured according to type; breath hydrogen (cyan), urine metabolites (yellow), Bifidobacterium (red), Dorea longicatena (purple), Blautia wexlerae (orange), Eubacterium hallii (brown), Lachnospiracaea (green), Anaerostipes (blue), Clostridiales (pink) and Unclassified (grey). m/z refers to the mass-to-charge ratio of a given unidentified urine metabolite. BAIBA β-aminoisobutyric acid, DHPPA 3,5-dihydroxy-hydrocinnamic acid
… 
This content is subject to copyright. Terms and conditions apply.
ARTICLE
A low-gluten diet induces changes in the intestinal
microbiome of healthy Danish adults
Lea B.S. Hansen et al.
#
Adherence to a low-gluten diet has become increasingly common in parts of the general
population. However, the effects of reducing gluten-rich food items including wheat, barley
and rye cereals in healthy adults are unclear. Here, we undertook a randomised, controlled,
cross-over trial involving 60 middle-aged Danish adults without known disorders with two 8-
week interventions comparing a low-gluten diet (2 g gluten per day) and a high-gluten diet
(18 g gluten per day), separated by a washout period of at least six weeks with habitual diet
(12 g gluten per day). We nd that, in comparison with a high-gluten diet, a low-gluten diet
induces moderate changes in the intestinal microbiome, reduces fasting and postprandial
hydrogen exhalation, and leads to improvements in self-reported bloating. These observa-
tions suggest that most of the effects of a low-gluten diet in non-coeliac adults may be driven
by qualitative changes in dietary bres.
DOI: 10.1038/s41467-018-07019-x OPEN
Correspondence and requests for materials should be addressed to R.G. (email: ramneek@bioinformatics.dtu.dk) or to T.R.L. (email: trli@food.dtu.dk)
or to O.P. (email: oluf@sund.ku.dk).
#
A full list of authors and their afiations appears at the end of the paper.
NATURE COMMUNICATIONS | ( 2018) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications 1
1234567890():,;
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Mechanistic and objective evidence on the effects of
excluding gluten-rich food items for healthy adults is
currently lacking, making the low-gluten diet highly
debatable in public. Although not the sole component changed in
a low-gluten diet, most discussion has centred on the dietary
component gluten. Gluten is a major dietary component in
wheat, rye and barley, and consists of proteins that are partially
resistant to proteolytic digestion due to a high content of proline
and glutamine1,2. Large gluten peptides including gliadin escape
gastric digestion and accumulate in the small intestine, where
they may interact with the immune system3,4, affect the intestinal
permeability57, and modify the gut microbial activity8,9. How-
ever, beyond the reduction in gluten, a low-gluten dietary regime
also entails a replacement of dietary bres of gluten-rich cereals
such as wheat, rye and barley with dietary bres from other
sources. Two short-term studies enroling 10 and 21 subjects
based upon 16S rRNA gene proling, respectively, have suggested
that a gluten-free diet (GFD) changes the gut microbiome and
immune function in healthy adults, however, with discrepant
results10,11. Thus, it remains unsettled if a low-gluten diet affects
the taxonomic and functional microbiome and host physiology of
healthy individuals. Here we report the results of a randomised,
controlled, cross-over trial encompassing 60 Danish adults
without coeliac disease. We nd that a low-gluten diet, in com-
parison with a high-gluten diet, induces changes in the compo-
sition and function of the gut microbiome (predened primary
outcome12), the urine metabolome and markers of host phy-
siology (Fig. 1a, b).
Results
Cross-over intervention. To examine the impact of a low-gluten
diet on the composition and function of the intestinal micro-
biome, urine metabolome and measures of host physiology, we
undertook a randomised, controlled, cross-over trial with two 8-
week dietary interventions comprising a low-gluten diet and a
high-gluten diet, separated by a washout period of at least
6 weeks12. The trial was conducted from July 2012 to November
201312. A total of 81 individuals were assessed for eligibility of
which 18 did not meet the inclusion criteria12 and three declined
to participate. Of notice, one excluded individual displayed ele-
vated serum transglutaminase concentration (a marker of coeliac
disease) and was excluded from the trial and referred for further
clinical investigation. Sixty Caucasian Danish adults without
coeliac disease, diabetes or any other self-known disorders were
included12. They were between 22 and 65 years old, healthy by
physical and biochemical examination, weight stable and had a
body mass index (BMI) of 2535 kg m2and/or increased waist
circumference (94 cm for men and 80 cm for women). No
study participants had a diagnosis of chronic disorders including
a gastrointestinal disease. Study participants were randomly
assigned to two groups: (1) undertaking either a low-gluten diet
followed by high-gluten diet, or (2) high-gluten diet followed by
low-gluten diet (Fig. 1a). In total, 51 participants completed the
study and 54 participants had more than two visits and were
included in the analyses (see baseline characteristics in Supple-
mentary Table 1 and CONSORT ow diagram in Supplementary
Fig. 1). During the two dietary interventions, study participants
were asked to replace all cereal products with freely provided low-
gluten or comparable gluten-rich dietary bre-matched products
of high-nutritional quality (Supplementary Table 2), which they
were asked to consume ad libitum.
Overall, participants were highly compliant to both interven-
tions, as documented in both food diaries (Supplementary
Table 3) and according to measured fasting plasma alkylresorci-
nol concentrations, which were substantially reduced on the low-
gluten diet compared with the high-gluten diet (Supplementary
Table 4; P< 0.001, linear mixed model), providing objective
evidence of individual compliance13. During the interventions,
study participants consumed on average ± standard deviation 2 ±
2 g gluten per day (mainly from oats) during the low-gluten
dieting period and 18 ± 6 g gluten per day (mainly from wheat
and rye) during the high-gluten dieting period, in comparison to
their habitual intake of 12 ± 4 g gluten per day (Supplementary
Table 5). The habitual intake of gluten is comparable with a mean
intake of 10.4 ± 4.4 g gluten per day in Denmark14, and the intake
of gluten in the low- and high-gluten diets are in line with a
previous study testing the effects of a low-gluten (2 g gluten
per day) and high-gluten (16 g gluten per day) diet in patients
with non-celiac gluten sensitivity15. Importantly, there was no
difference between the two diet regimens in intake of total dietary
bre content. Intake of wholegrain cereals (wheat, rye and barley)
was as expected lower in the low-gluten diet compared with the
high-gluten diet (Supplementary Table 5; P< 0.001, paired t-test).
There were no differences between the interventions in total
energy or macronutrients intake, except for a slightly reduced
protein intake during the low-gluten period (on average reduced
with 7 g per day during the low-gluten period; Supplementary
Table 5; P=0.01, paired t-test). We compared the effects of the
diets on changes in composition and functional potential of the
gut microbiome, the urine metabolome, targeted serum and
faeces metabolites and markers of host physiology using
measurements of each variable taken at baselines (visit 1 and
visit 3) and at end-points (visit 2 and visit 4) (Fig. 1a, b).
A low-gluten diet alters the intestinal microbiome. To estimate
a potential impact of low-gluten versus high-gluten dieting on the
gut microbiome, we studied a total of 208 individual whole-
genome shotgun sequences of microbial DNA obtained from
stool samples. On average, we obtained 6.7 Giga base-pairs (bp)
per sample when including samples ranging from 3.7 to 13.6 Gbp
(Supplementary Data 1). The microbial sequences were mapped
to the integrated catalogue of reference genes of the human gut
microbiome16 and genes were binned into metagenomic species
(MGS; informal distinct microbial entities, from hereon called
species) according to co-abundance variation across samples17.In
total, 575 species were identied in at least ten individuals in this
cohort. Of these species, the relative abundance of 14 bacterial
species was altered during the low-gluten diet intervention
compared with the high-gluten diet intervention (Fig. 2and
Supplementary Data 2; false-discovery rate (FDR) < 0.05, linear
mixed model). Consistently, the abundance of four species of
Bidobacterium was diminished during the low-gluten diet
(Supplementary Fig. 2). The substantial reduction in Bido-
bacterium spp., both in terms of absolute and relative abundance,
were conrmed by quantitative PCR (Supplementary Table 6). In
addition, the low-gluten diet resulted in a decrease of a species
annotated as Dorea longicatena and another species of Dorea, one
species of Blautia wexlerae, two species of the Lachnospiraceae
family, and two butyrate-producing bacteria Anaeostipes hadrus
and Eubacterium hallii, in comparison with the high-gluten diet.
At the same time, an unclassied species of unknown taxonomic
origin, an unclassied species of Clostridiales and an unclassied
species of Lachnospiraceae increased during the low-gluten diet
intervention compared with the high-gluten diet intervention.
Notably, we did not nd any changes in alpha- and beta-diversity
(Supplementary Fig. 3).
To explore changes in the functional capacity of the intestinal
microbiome following low-gluten as compared with the high-
gluten dieting, all microbial genes annotated to prokaryotic Kyoto
Encyclopedia of Genes and Genomes (KEGG) orthologies (KOs)
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x
2NATURE COMMUNICATIONS | (201 8) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
18 were tested individually, when grouped into KEGG modules19
and when grouped into customised reference
modules20, respectively. We identied 88 KOs and 37 modules
that changed following the low-gluten diet period compared with
the high-gluten diet intervention (Fig. 3and Supplementary
Data 3 and 4; FDR < 0.05, linear mixed model). In particular, the
abundance of modules associated with carbohydrate metabolism
(i.e. arabinose degradation, pentose phosphate pathway, phos-
phate acetyltransferase-acetate kinase pathway and fructose-6-
phosphate shunt) and uptake of carbohydrates (L-arabinose/
lactose transport system, phosphotransferase systems (PTS) and
other sugar transport systems) was diminished following the low-
gluten dieting compared with the high-gluten dieting. This
suggests a change in bacterial carbohydrate degradation as a
response to the dietary intervention. Furthermore, abundance of
modules associated with bacterial transport of glutamate, zinc/
manganese and sulphate was diminished, whereas abundance of
modules associated with transport of cysteine and iron was
increased following the low-gluten diet compared with the high-
gluten diet period. A majority of the modules showed subtle
changes in the functional potential (Fig. 3a, b), high prevalence in
the species (Fig. 3c) and the signicantly different species
comprised a minor percentage of the total functional poten-
tial (Fig. 3d). However, leaving out the signicantly different
species showed that they contributed considerably to the observed
signicant changes in functional potential for multiple of the
Low-gluten diet
Visit 1 Visit 2 Visit 3 Visit 4
8 weeks 6 weeks
High-gluten diet
Low-gluten diet
High-gluten diet
Washout
60 subjects
Intestinal microbiome
Anthropometrics
Blood tests
Gut permeability
Untargeted urine metabolomics
Targeted faeces & serum metabolomics
Standardized meal test
Breath hydrogen
Gastrointestinal symptoms
Intestinal transit time
4 days dietary records
a
b
Differences in nutrient content including
changes in gluten components and dietary fibres
Alterations in bacterial
species (MGSs):
Bifidobacterium
Eubacterium halli
Anaerostipes hadrus
Blautia wexlerae
Dorea
Changes in functional potential:
Carbohydrate degradation
Sugar transport
Urine metabolome:
Fibre derived metabolites
Faecal metabolome:
Kynurenine
SCFA
Host physiology:
Breath hydrogen
Bloating
Post-meal PYY
Body weight
Reduced following a low-gluten diet
Increased following a low-gluten diet
No difference between the diets
Low-gluten diet vs. High-gluten diet
8 weeks
NH2
NH2
O
OH H3C
H3C
O
O
O
O
O
O
O
O
O
O
O
O
O
NH2
NH2
OH
Fig. 1 Experimental design, data overview and summary of the cross-over trial. aThe study was a randomised, controlled, cross-over trial with two 8-week
dietary intervention periods separated by a washout period of at least six weeks, comparing the effects of a low-gluten diet and a high-gluten diet on the
gut microbiome (predened primary outcome), untargeted urine metabolome and measures of host physiology12. Time points for data collections are
indicated by circles in the lower part panel (a). bEffects of a low-gluten diet compared with a high-gluten diet on the intestinal microbiome, urine/faecal
metabolome and markers of host physiology in apparently healthy adults. Measured variables that were found to be reduced (red arrow), increased (green
arrow) or unchanged (black horizontal arrows) following the low-gluten diet intervention compared with the high-gluten diet intervention are listed. MGS
metagenomics species, PYY peptide YY, SCFA short-chain fatty acids. The person icon and molecular structure images for the acetate anion, butyrate ion,
propionate ion and kynurenine were obtained from Wikimedia Commons, released under public domain
NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x ARTICLE
NATURE COMMUNICATIONS | ( 2018) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved
modules (Fig. 3b). In summary, these ndings demonstrate that
low-gluten dieting changes the gut microbiome composition and
functional potential in healthy adults.
A low-gluten diet changes the intestinal fermentation.We
found a reduction in both fasting and postprandial hydrogen
exhalation after an identical standardised test meal following low-
gluten dieting compared with high-gluten diet dieting (Fig. 4a and
Supplementary Table 7; P<0.0001, linear mixed model). In
addition, participants reported improved postprandial well-being
after the standardised meal following the low-gluten diet com-
pared to the high-gluten diet (Supplementary Fig. 4). Whereas the
Adjusted P value
B. angulatum
B. longum
B. adolescentis
B. pseudocatenulatum
D. longicatena
Dorea
Lachnospiraceae 1
Lachnospiraceae 2
B. wexlerae
E. hallii
Unclassified 1
A. hadrus
Unclassified 2
Clostridiales
Effect size - log10
Unclassified 1
Clostridiales
Lachnospiraceae 2
Average Log2FC
Increase
Decrease
Adjusted P value
Total abundance (%)
Subjects
42
11
35
48
12
51
50
30
51
47
47
33
33
17
Increased
Low-gluten diet High-gluten diet
5e–2 1e–4 1e–10
0.1 0.5 1
Total abundance (%)
–5 5
–2 2
–3 3
–4 4
B. adolescentis
B. angulatum
E. hallii
B. pseudocatenulatum
A. hadrus
Unclassified 2
Lachnospiraceae 1
Dorea
D. longicatena
B. longum
B. wexlerae
a
b
Adjusted P value
Bacteroidaceae
Bifidobacteriaceae
Clostridiaceae
Decreased in the low–gluten diet compared to the high–gluten diet
Lachnospiraceae Porphyromonadaceae
Oscillospiraceae
Other
Rikenellaceae
Ruminococcaceae
Total abundance (%) 12
0.40.0–0.4
1e–06
1e–03
1e+00
Eubacteriaceae
Erysipelotrichaceae
Family
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x
4NATURE COMMUNICATIONS | (201 8) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
change in breath hydrogen was primarily driven by the low-
gluten diet, the change in postprandial well-being was unex-
pectedly primarily driven by the high-gluten diet. However, the
reduction in breath hydrogen was convergent with participants
reporting less bloating following the 8-week low-gluten inter-
vention compared with the 8-week high-gluten intervention
(Fig. 4b). Together these observations suggest an altered intestinal
fermentation in accordance with the changes in bacterial modules
associated with carbohydrate metabolism (Fig. 3). Indeed, several
differences in carbohydrate composition were found between
diets including higher levels of galactose, rhamnose, mannose,
and galacturonic acid and lower levels of arabinose and xylose in
the low-gluten diet compared with the high-gluten diet
(Supplementary Fig. 5a). These nutritional changes were in
agreement with a reduced bacterial arabinose/lactose transport
potential following the low-gluten dieting (Fig. 3). There was no
differences in the total amount of dietary fermentable, oligo-, di-,
and monosaccharides and polyols (FODMAP) (Supplementary
Fig. 5b) or in intake of resistant starch (Supplementary Table 2)
between the two diets. Yet, qualitative differences were observed,
such as lower levels of fructooligosaccharides and mannitol/sor-
bitol and higher levels of lactose in the low-gluten diet. In support
of a changed intestinal fermentation, breath hydrogen con-
centrations were negatively associated with gut metabolic mod-
ules related to methanogenesis (Supplementary Table 8). The
latter comprises reduction of CO
2
to CH
4
using H
2
or formate as
Custom modules
decreased during
low-gluten diet
Quinolinic acid synthesis
Fructose-6-phosphate shunt
Arabinose degradation
Sucrose degradation l
Acetyl-CoA to acetate
Pentose phosphate pathway (non-oxidative branch)
Diaminopimelic acid (DAP) & MurJ flippase (Gram-negative peptidoglycan)
Sucrose degradation ll
Glutamate transport system
PTS system, mannose-specific ll component
PTS system, glucitol/sorbitol-specific ll component
Multiple sugar transport system
Bacterial proteasome
F-type ATPase, prokaryotes and chloroplasts
Nucleotide sugar biosynthesis, glucose => UDP-glucose
Fatty acid biosynthesis, initiation
Putative zinc/manganese transport system
PTS system, cellobiose-specific ll component
Multiple sugar transport system
Methyl-galactoside transport system
PTS system, N-acetylgalactosamine-specific ll component
PTS system, sucrose-specific ll component
Holo-TFIIH complex
Shikimate pathway, phosphoenolpyruvate + erythrose-4P => chorismate
Sulfate transport system
L-Arabinose/lactose transport system
Putative fructooligosaccharide transport system
PTS system, lactose-specific ll component
Glycolysis (Embden-Meyerhof pathway), glucose => pyruvate
C10-C20 isoprenoid biosynthesis, archaea
Cystine transport system
Methionine salvage pathway
Fatty acid biosynthesis, elongation
Iron complex transport system
Pentose phosphate pathway, non-oxidative phase, fructose 6P => ribose 5P
Phosphate acetyltransferase-acetate kinase pathway, acetyl-CoA => acetate
KEGG modules
decreased during
low-gluten diet
KEGG modules
increased during
low-gluten diet
–1.0 –0.5
abcd
Low-gluten diet
High-gluten diet
All MGSs
Significant MGSs left out
0.0
Log2FC of median abundance Adjusted P value (–log10) MGS prevalence % of module abundance by
significantly differing MGSs
0.5 0 5
p = 0.05
10 15 0 500 1000 0 20 40 60
B. longum
B. angulatum
B. pseudocatenulatum
B. adolescentis
A. hadrus
E. halli
B. wexlera
D. longicatena
Dorea
Unclassified 2
Lachnospiraceae 1
Lachnospiracaea 2
Clostridiales
Unclassified 1
Fig. 3 A low-gluten diet alters the functional potential of the gut microbiome. aMicrobial genes annotated to Kyoto Encyclopedia of Genes and Genomes
(KEGG) orthologs (KOs) were grouped into KEGG modules and manually curated (customised) modules. The bar chart display the median log2 fold
change (median ± SEM) of all individual KOs within a module when comparing the relative abundance at baseline to the abundance following the low-
gluten diet (blue bars) or the high-gluten diet (red bars), respectively. bDot plot of the negative log10 of the adjusted Pvalue from the linear mixed model
comparing changes in the abundance of modules induced by the low-gluten diet with the changes induced in the high-gluten diet (black dots) adjusting for
age, gender, intestinal transit time, participant (n=51) and carry-over effect. The same analysis was carried out while removing the signicant MGSs from
the data (grey dots) to elucidate their contribution to the signicance. All effect sizes and SEM for each KO can be found in Supplementary Data 3 and 4. c
Prevalence of the module across the 1264 MGSs identied from the IGC catalogue16,17. A module was assessed to be present or partially present in a MGS
when at least two KOs from the module were detected in the MGS.dBar plot showing the fraction of the total abundance of a module contributed by each
signicantly different MGS in per cent. (Supplementary Data 2)
Fig. 2 A low-gluten diet alters the composition of the gut microbiome. aScatterplot of the statistical signicance of the metagenomic species (MGSs) as
assessed by a linear mixed model testing for the difference between the low-gluten and the high-gluten diets adjusted for age, gender, intestinal transit
time, participant (n=51) and carry-over effect. Adjusted Pvalues are displayed on the y-axis (log10 scale) and the effect size (absolute values were log10
transformed) is on the x-axis. Points are sized according to the total abundance (%) and coloured according to the ten most abundant taxonomic families.
The Othercategory consists of the remaining families. The horizontal line represents an adjusted Pvalue of 0.05 and the 14 species that changed
signicantly (FDR < 0.05) between the interventions are labelled with their full taxonomic annotation. Only species that could be annotated to family level
and with abundance above 0.02% were included in the plot (255 species). bBar chart of the 14 signicant species showing the log2 fold change (means ±
SEM) between baseline and after the low-gluten diet (blue bars) and high-gluten diet (red bars), respectively. The black circles are sized according to the
negative log10 of the adjusted Pvalues of comparison between the low-gluten and the high-gluten diet using a linear mixed model adjusted for age, gender,
intestinal transit time, participant (n=51) and carry-over effect. Green circles are scaled according the species abundance. The last column lists the
number of participants in whom the given species were measured. Details on the individual species can be found in Supplementary Data 2
NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x ARTICLE
NATURE COMMUNICATIONS | ( 2018) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved
electron donors21 occurring alongside proteolytic degradation
following prolonged intestinal transit times22. We did not nd,
however, any changes in intestinal transit time (Supplementary
Table 7).
To further explore changes in intestinal fermentation, we
performed untargeted metabolic proling of urine sampled
during the standardized meal tests by gas chromatography mass
spectrometry (GC-MS) as well as by ultra-performance liquid
chromatography mass spectrometry (UPLC-MS). We found
lower concentrations of wheat-derived compounds (3,5-dihy-
droxyhydrocinnamic acid-glucuronide and galactosylglycerol)
during the low-gluten intervention in comparison with the
high-gluten intervention. In contrast the urinary concentration of
a host-microbial co-metabolite of lignan degradation (enterolac-
tone-glucuronide) was increased during the low-gluten interven-
tion (Supplementary Table 9; FDR < 0.05, linear mixed model),
suggesting an altered dietary bre degradation upon reduction in
gluten-rich food items during the low-gluten diet and related
changes in the gut microbiome.
To identify correlations between breath hydrogen levels and
changes in the gut microbiome and urine metabolome, we
developed a co-occurrence network of breath hydrogen and the
bacterial species and urine metabolites that responded to the
dietary interventions (Fig. 4c and Supplementary Data 5). Breath
hydrogen was positively associated with the wheat-associated
urine metabolites and B. longum and negatively associated with
urine enterolactone-glucuronide, substantiating that differences
in composition of dietary bres between the two diet regimens
resulted in a changed colonic fermentation. The network analysis
identied the lactate-utilising, butyrate-producing Eubacterium
hallii as a key driver species, which was associated with the
lactate-producing Bidobacterium, as well as the hydrogen-
producing Dorea longicatena and the hydrogen-consuming,
acetate-producing Blautia. Furthermore, the network analysis
highlighted associations between the wheat-associated urine
metabolites and the Bidobacterium species, suggesting that the
reduction in Bidobacterium abundance following the low-gluten
diet intervention was associated with the diminished intake of
wheat. Likewise, the abundance of microbiome modules asso-
ciated with uptake and degradation of mannose, sucrose and
arabinose was positively associated with wheat-associated urine
metabolites (Supplementary Fig. 6 and Supplementary Table 10).
Collectively, these results suggest that a changed gut microbiome
and altered fermentation resulting from qualitative differences in
Time (min)
Low-gluten diet
High-gluten diet
0
Baseline
30
60
90
120
180
150
30
20
0
10
40 100
50
–50
–100
mz197.0558
mz222.0451
Galactosylglycerol
mz226.1119
mz179.0035
Phenol metabolite
mz206.0116
B. wexlerae
Cystathionine
N-Acetyl-
Cystathionine
Clostridiales
Unclassified 1
Unclassified 2
Lachnospiraceae 2
B. adolescentis
E. haiiii
A. hadrus
B. pseudo-
catenulatum
D. longicatena Lachnospiracea 1
Dorea
B. angulatum
B. longum
BAIBA
Guanidino-
succinic acid
2,3-Diamino-
salicylic acid
Enterolactone-
glucuronide
DHPPA-glucuronide
Breath hydrogen
0
*
*
*
**
**
Post-meal breath hydrogen
(ppm)
Low-gluten diet start
Low-gluten diet end
High-gluten diet start
High-gluten diet end
Bifidobacterium
Dorea
Blautia
Eubacterium
Clostridiales
IncreaseDecrease
Low-gluten compared with a high-gluten diet
Unclassified
10–14
10–10
10–6
10–2
Breath hydrogen
Adjusted P value
Urine metabolites
Lachnospiraceae
Anaerostipes
ΔGut bloating
(VAS)
**
a
c
b
Fig. 4 A low-gluten diet affects measures of intestinal fermentation. aBreath hydrogen levels following the same standardised meal at all four visits (low-
gluten diet start, open blue circles; low-gluten diet end, blue squares; high-gluten diet start, open red triangle; high-gluten diet end, lled red triangle). Data
are shown as means ± SEM (n =51-57). bPlot showing changes in gut bloating as assessed by visual analogue scale (VAS) following the low-gluten diet
(blue circles) compared with the high-gluten diet (red squares). Data are shown as means ± SEM (n=5253). Changes were assessed by a linear mixed
model adjusting for age, gender and intestinal transit time. *P< 0.05, **P< 0.01. cLinear regression network of breath hydrogen levels and the abundance
of bacterial species and concentrations of urine metabolites which are signicantly responding to the dietary interventions using a linear mixed model
adjusted for gender, age and participant (n=49) (Supplementary Data 5). The dotted line separates the features that were decreased and increased,
respectively, when comparing the low-gluten and high-gluten periods. Signicant (FDR < 0.05) positive associations are indicated with grey lines; negative
associations with red lines. Thickness of lines indicates the signicance level. Nodes are coloured according to type; breath hydrogen (cyan), urine
metabolites (yellow), Bidobacterium (red), Dorea longicatena (purple), Blautia wexlerae (orange), Eubacterium hallii (brown), Lachnospiracaea (green),
Anaerostipes (blue), Clostridiales (pink) and Unclassied (grey). m/z refers to the mass-to-charge ratio of a given unidentied urine metabolite. BAIBA β-
aminoisobutyric acid, DHPPA 3,5-dihydroxy-hydrocinnamic acid
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x
6NATURE COMMUNICATIONS | (201 8) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
dietary bre composition may explain the reduced breath
hydrogen and reduced bloating following the low-gluten diet.
A low-gluten diet results in weight loss. We did not nd any
differences in measures of glucose and lipid metabolism (Sup-
plementary Table 7). However, we found a decrease in body
weight, on average 0.8 ± 0.3 kg, following the low-gluten dieting
for 8-week compared with the high-gluten diet period (Fig. 5a;
P=0.012, linear mixed model). We also demonstrated an
increase in postprandial plasma concentrations of peptide YY
(PYY) in response to the standardised meal after the low-gluten
intervention compared with the high-gluten intervention (Fig. 5b
and Supplementary Table 7; P
AUC
=0.012, linear mixed model).
PYY is a gut hormone released into the circulation in a nutrient-
dependent manner and is known to reduce appetite23. However,
we did not observe any differences in total energy intake during
the two interventions (Supplementary Table 5). There were no
changes in levels of the proximal incretin hormone gastric inhi-
bitory peptide (GIP), and no changes in the distal gut hormone
glucagon-like peptide 2 (GLP-2), a regulator of gut mucosal
adaptation and growth; GLP-2 is known to be secreted in com-
parable amounts in parallel to the appetite regulating hormone
GLP-1, and it may therefore be assumed that also GLP-1 secre-
tion was unchanged24 (Supplementary Table 7). Together, these
ndings suggest that a different mechanism is responsible for the
weight loss. Colonic short-chain fatty acids (SCFA), synthesized
by the gut microbiota during bre fermentation, are known to
increase plasma PYY levels, fat oxidation and energy expenditure
in overweight men25. However, we did not observe any associa-
tions between changes in fasting or postprandial plasma PYY
concentrations and changes in bacterial modules associated with
SCFA biosynthesis potentials (Supplementary Table 8) or in
faecal and serum concentrations of SCFA (Supplementary
Table 11). Fasting plasma PYY concentrations have been nega-
tively associated with various markers of adiposity and resting
metabolic rate in humans26, and long-term elevated PYY con-
centrations are associated with enhanced thermogenesis in
mice27. Among the differing urine metabolites, β-aminoisobutyric
acid (BAIBA) was increased following the low-gluten diet com-
pared with the high-gluten diet period (Fig. 5c and Supplemen-
tary Table 9; FDR < 0.05, linear mixed model). BAIBA induces
browning of white adipose tissue and increases hepatic fat
oxidation28. Changes in urine BAIBA concentrations were,
however, not associated with the bacterial pyrimidine degradation
module (Supplementary Table 8), which contains bacterial genes
involved in degradation of thymine into BAIBA, suggesting that
the observed changes in urine BAIBA were not related to changes
in the intestinal microbiome. Rather they might be directly
related to other effects of the low-gluten diet on host metabolism.
Together, the increased urine concentrations of BAIBA and the
elevated postprandial plasma levels of PYY suggest that intake of
the low-gluten diet modulated energy homoeostasis by changing
thermogenesis or fat oxidation. To explore these hypotheses, we
performed targeted metabolomics quantifying fatty acids, acyl-
carnitines (transport fatty acids into the mitochondria for
breakdown), and BAIBA in serum. Besides a signicant increase
in serum linoleyl-carnitine following the low-gluten diet com-
pared with the high-gluten diet, these metabolites were not
changed (Supplementary Table 12), suggesting unaltered fat
oxidation. Further exploring possible reasons for the observed
weight loss, we targeted metabolites associated with the
microbiota-gut-brain axis including serotonin, kynurenine, glu-
tamate, γ-aminobutyric acid29,30 in faeces and serum. Analyses of
these metabolites did reveal a signicant increase in faecal
kynurenine concentrations following the low-gluten diet inter-
vention compared with the high-gluten diet intervention (Fig. 5d
and Supplementary Table 12; P=0.005, linear mixed model).
Compared with healthy controls, coeliac disease patients adhering
to a GFD have been reported to have lower serum concentrations
of aromatic amino acids including tryptophan, the substrate for
kynurenine31. Therefore, we quantied serum and faeces con-
centrations of the aromatic amino acids and their derivatives.
Since concentrations of tryptophan and microbial tryptophan
catabolites were unaltered (Supplementary Table 12), the
observed increase in kynurenine faeces concentration following
the low-gluten diet suggested altered microbiota tryptophan
degradation pathways rather than being a mere consequence of
substrate availability. Indeed, targeted metagenomic module
analyses revealed a proportional decrease in the potential of the
tryptophan to serotonin synthesis pathway (Supplementary
Data 6 and Supplementary Fig. 7) following the low-gluten diet.
Moreover, we found the ratios of the proportional abundances of
their respective production pathways and faecal concentrations to
correlate (Supplementary Fig. 8a; Spearman rho =0.20; P=
0.004; Supplementary Table 13), suggesting a balance between
both tryptophan conversion routes (Supplementary Fig. 7).In
rodents, kynurenic acid, a downstream product of kynurenine,
Low-gluten diet
High-gluten diet
Low-gluten diet
High-gluten diet
Low-gluten diet
High-gluten diet
–6
–4
–2
0
2
4
6
8
ΔBody weight (kg)
*
–3
–2
–1
0
1
2
3
Urine BAIBA
(Log2 fold change)
***
0
30
60
90
120
180
8
10
12
14
16
18
Time (min)
Plasma PYY (pM)
Low-gluten diet start
Low-gluten diet end
High-gluten diet start
High-gluten diet end
**
PAUC=0.012
0
–5
0
5
10
Faecal Kynurenine
(Log2 fold change)
**
ab c d
150
Fig. 5 Low-gluten dieting affects markers of host metabolism. aPlot showing participantschanges in body weight following the low-gluten (blue circles)
and high-gluten (red squares) periods. bPlot showing participantsplasma concentrations of peptide YY (PYY) following a standardised meal at all four
visits (low-gluten diet start, open blue circles; low-gluten diet end, blue squares; high-gluten diet start, open red triangle; high-gluten diet end, lled red
triangle). cPlot showing log2 fold changes in participantsurine concentrations of β-aminoisobutyric acid (BAIBA) and dfaecal concentrations of
kynurenine following the low-gluten (blue circles) and high-gluten (red squares) diet, respectively. Data are shown as means ± SEM, n=5054. Changes
were assessed by a linear mixed model adjusting for age, gender and intestinal transit time. *P< 0.05, **P< 0.01, ***P< 0.001. AUC area under the curve
NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x ARTICLE
NATURE COMMUNICATIONS | ( 2018) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved
has been reported to enhance adipose tissue thermogenesis
through activation of G protein-coupled receptor Gpr3532, which
is also highly expressed in the gastrointestinal tract33. Here, we
observed faecal kynurenine concentrations to be positively asso-
ciated with urine BAIBA levels (Supplementary Fig. 8b; Spearman
rho =0.26; P=9.2E05), indicating a potential role of the colon
microbial production in fat browning.
A low-gluten diet has subtle effects on the immune system.To
determine the potential impact of a low-gluten diet on immune
and inammatory host responses, we assessed systemic inam-
matory markers as well as ex vivo lipopolysaccharide (LPS)-
induced cytokine responses in whole-blood of study participants.
We did not nd any changes in concentrations of systemic
inammatory markers in serum (i.e., C-reactive protein (CRP),
interleukin (IL)-6 or tumour necrosis factor alpha (TNF-α);
Supplementary Table 7) or in counts of immune cell populations
in blood (i.e. leucocytes, lymphocytes, neutrophils, monocytes;
Supplementary Table 7). Neither did we nd any changes in
markers of intestinal inammation (i.e. fasting plasma citrulline
and faecal calprotectin; Supplementary Table 7) nor in intestinal
permeability as measured by fasting serum zonulin and urinary
excretion of lactulose and mannitol (Supplementary Table 7). Of
notice, ex vivo LPS-induced stimulation of whole-blood showed,
however, reduced release of the pro-inammatory,
inammasome-related cytokine IL-1βfollowing the low-gluten
diet intervention compared with the high-gluten diet period
(Supplementary Table 7; P=0.035, linear mixed model). None of
the other serum concentrations of non-inammasome-related,
pro-inammatory mediators such as IL-6, TNF-αand interferon
gamma (IFN-γ) were changed. These ndings suggest a selec-
tively reduced activation of the inammasome response following
the low-gluten diet intervention compared with the high-gluten
diet period. Intriguingly, we demonstrated a positive association
between the abundance of the bacterial Lipid A synthesis module
(present in all Gram-negative bacteria) and LPS-induced release
of IL-1βfrom whole-blood (Supplementary Table 8). Collectively,
these results suggest that a low-gluten diet confers a selectively
reduced activation of the inammasome response.
Discussion
An overview of the outcome of this randomised, controlled,
cross-over trial with two 8-week dietary intervention periods
comparing the effects of a low-gluten diet and a high-gluten diet
is given in Fig. 1.
We showed that a low-gluten diet in apparently healthy adults
changed the primary trial endpoint, the gut microbiome com-
position and functional potential. Among the 14 bacterial species
which changed between the two dietary regimens, particularly the
relative abundance of Bidobacterium species was consistently
diminished following adherence to the low-gluten dietary regi-
men. This nding is in agreement with a microbiota gene marker
study involving 10 healthy adults showing that a shift to a GFD
for four weeks resulted in decreased proportions of Bido-
bacterium10, as well as with reports of lower abundance of bi-
dobacteria in celiac disease patients following a GFD34,35.In
addition, practicing a low FODMAP diet diminishes the abun-
dance of bidobacteria in patients with IBS concurrent with relief
of gastrointestinal symptoms3638. These interventions generally
reduce intake of wheat or exclude wheat, suggesting a close
relationship between wheat intake and the abundance of bido-
bacteria in adults. This aligns with recent studies showing that
healthy populations living traditional lifestyles have low or absent
faecal abundance of bidobacteria compared with the intestinal
ecosystems of individuals in industrialised parts of the world39,40.
Thus, the abundance of bidobacteria in adults living a Western
lifestyle may to a large extent reect intake of diets enriched in
wheat.
In parallel, we observed a reduction in butyrate-producing E.
hallii and A. hadrus as well as in the hydrogen-producing Dorea
and the hydrogen-consuming, acetate-producing Blautia, fol-
lowing the low-gluten diet compared with the high-gluten diet.
These interrelated species were positively associated, which is
consistent with reports on cross-feeding between Bidobacterium
and butyrate-producing bacteria4143, with Blautiaability to
produce acetate and utilise hydrogen during bre
fermentation41,44,45 and with the ability of Dorea longicatena to
produce hydrogen46. Several in vivo and in vitro studies have
shown bidogenic effects and stimulation of butyrate-producing
colon bacteria4751 by arabinoxylan and arabinoxylan-oligo-
saccharides, abundant non-starch polysaccharides of cereal
grains52. Indeed, the bre composition analysis of the two
intervention diets showed lower concentration of arabinose in the
low-gluten products compared to the high-gluten products. This
was in agreement with changes in the functional potential of the
microbiome upon the low-gluten dieting. A module representing
a L-arabinose/lactose transport system and a custom module
representing arabinose degradation, which converts L-arabinose
to L-ribulose-5-phosphate, were signicantly reduced during a
low-gluten diet. Further, L-ribulose-5-phosphate is utilized by the
non-oxidative phase of the pentose phosphate pathway, which
was also signicantly reduced during the low-gluten period
compared to the high-gluten period. This suggests that the
replacement of grain-derived bres of wheat, barley and rye with
dietary bres of other sources during the low-gluten diet inter-
vention caused the observed changes in the intestinal micro-
biome. Importantly, in accordance with the diets being matched
for dietary bres, we did not during our 8-week intervention
observe any changes in faecal and serum SCFA. Furthermore, we
did not nd any health implications associated with the reduction
in Bidobacterium and butyrate-producing species following the
low-gluten diet, although the long-term health consequences
remain unknown.
Despite the unchanged concentrations of SCFA concentration
in serum and faeces, we observed a reduction in both fasting and
postprandial hydrogen exhalation following the low-gluten diet
intervention and multiple changes in urine metabolites reecting
a changed intestinal fermentation. In line with this, a previous
study reported that fasting breath hydrogen concentrations were
signicantly lower in coeliac disease patients on a GFD compared
with untreated coeliac disease patients53. Likewise, a low FOD-
MAP diet has been reported to reduce breath hydrogen and
ameliorate gastrointestinal symptoms compared with a high
FODMAP diet15,37. To which extent the concurrent improve-
ments in well-being and bloating following the low-gluten diet
intervention as compared with the high-gluten diet period, were
prompted by changes in the intestinal microbiome and fermen-
tation, or were due to psychological (placebo) effects remain
unresolved.
No effects on glucose and lipid metabolism were found.
However, despite unaltered self-reported energy intake by study
participants, the low-gluten diet was temporarily linked with a
signicant weight loss. This is in line with two studies in mice fed
a gliadin-enriched, high-fat diet showing an increase in body
weight and adiposity54,55. Still, other studies in mice show no
effect of gluten on body weight3,56,57. Based on the observed
increase of plasma PYY and urinary BAIBA concentrations fol-
lowing the low-gluten diet, we hypothesized that the reduced
body weight induced by low-gluten intervention might in part be
mediated by an increased thermogenesis. Recent studies in mice
have indicated that increased intake of gluten may increase
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x
8NATURE COMMUNICATIONS | (201 8) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
hepatic lipid accumulation3, reduce the thermogenic capacity of
adipose tissue55 and the size of adipocytes3. Here we found that
faecal concentrations of kynurenine were increased following the
low-gluten diet and associated with urine BAIBA, raising the
intriguing possibility that kynurenine, via the downstream pro-
duct of kynurenic acid, enhance thermogenesis32 through acti-
vation of Gpr35 in the gastrointestinal tract33. Obviously, further
interventions in humans are warranted to specically delineate
whether intake of a low-gluten diet modulates energy
homoeostasis.
Our analyses showed that the low-gluten diet had no effect on
circulating white blood cell counts or markers of systemic
inammation in unstimulated blood or on measures of intestinal
inammation. Likewise, no effects were seen on intestinal per-
meability markers. We did, however, notice that LPS-induced
stimulation of whole-blood showed that immune cells had
reduced capacity to produce the pro-inammatory,
inammasome-related cytokine IL-1β58 following the low-gluten
diet period. Similarly, a previous study reported that production
of pro-inammatory cytokines by peripheral blood mononuclear
cells stimulated with faecal water was reduced after a GFD10.As
the inammasome-directed response also takes place in intestinal
cells in a similar manner, and is regarded as an important reg-
ulator of intestinal homoeostasis59,60, these ndings might point
to a yet undescribed impact of a low-gluten diet on the immune
system that in future studies will need further clarication.
In conclusion, an 8-week low-gluten diet intervention in
healthy middle-aged adults induced changes in the intestinal
microbiome and fermentation of complex carbohydrates as
mirrored in changes of the urine metabolome and reduction in
breath hydrogen. Although the generalizability to other popula-
tions is to be determined as gluten consumption differs in Wes-
tern populations61,62, the changes in colonic microbial
composition and fermentation suggest that the effects of a low-
gluten diet in healthy middle-aged adults may to some extent be
driven by qualitative changes in dietary bres upon reduction of
gluten-rich food items rather than by the reduction of gluten
intake itself.
Methods
Trial design. This was a randomised controlled (1:1) cross-over trial composed of
two 8-week dietary interventions comprising a low-gluten diet or a high-gluten
diet, separated by washout period for at least six weeks (range 623 weeks, median
of 8 weeks) with habitual diet. The trial design, intervention modes and primary
and secondary outcomes have been reported in a previous paper12 and registered at
www.clinicaltrials.gov (NCT01719913). The trial was conducted from July 2012 to
November 201312.
Participants. Participants were recruited from the general population studies
Health 2008and Health 2010, established at the Research Center for Preven-
tion and Health (RCPH) at Glostrup University Hospital in Copenhagen, Den-
mark63 and through the webpage www.forsogsperson.dk and advertisements in
local newspapers. Participants were non-diabetic, lean, overweight or obese adults
who were healthy by self-report and aged 2265 years. Importantly, they did not
suffer from coeliac disease or other gastrointestinal diseases. In order to detect
latent coeliac diseases, levels of serum Immunoglobulin(Ig)A and IgG transgluta-
minase were measured at the rst examination day. In case values exceeded the
acceptable maximum (> 8 units per mL for IgA and >10 units per mL for IgG)
participants were excluded from the study and referred to own general practitioner.
Further eligibility criteria have been published elsewhere12. Exclusion criteria
included antibiotic treatments (< 3 months prior to study start), intake of pre- or
probiotic supplements (<1 month prior to study start), medically prescribed diet
and intense physical activity (>10 h per week)12. Data on participantsphysiological
traits and smoking habits are available in Supplementary Table 1.
The study was led by the Novo Nordisk Foundation Center for Basic Metabolic
Research, Faculty of Health and Medical Science, University of Copenhagen and
conducted at the Department of Nutrition, Exercise and Sports at the University of
Copenhagen, Denmark. The Ethical Committee of the Capital Region of Denmark
approved the trial (H-2-2012-065), which was run in accordance with the Helsinki
declaration and endorsed by the Data Protection Agency (2007-54-0269). All
individuals gave written informed consent before participating in the study.
Interventions. The aim of the dietary interventions was to limit the daily gluten
consumption considerably in the low-gluten period (~2 g per day) and to increase
it in the high-gluten period (~20 g per day). For comparison, in the national survey
of dietary habits, Danish adults (n=1494, 2075 years) had a mean total gluten
intake of 12.0 ± 4.6 g per day in men and 9.0 ±3.4 g per day in women14. During
the two dietary interventions participants were provided with a selection of low-
gluten or high-gluten products of high nutritional values and instructed to replace
all cereal products from their habitual diet with the study dietary products and to
consume these products ad libitum (Supplementary Table 2). Each participant was
randomly assigned to start on either the low-gluten diet or the high-gluten diet.
Participants were encouraged to contact the study staff if they experienced any
adverse health-related implications of the dietary interventions. The outline of the
trial is shown in Fig. 1.
Overview of protocol measures. The primary endpoint was an altered gut
microbiota composition and functional potential during consumption of a low-
gluten compared with a high-gluten diet as measured by shotgun sequencing-based
metagenomics analyses of microbial DNA isolated from faecal samples and
sequenced applying deep metagenomics sequencing12. Secondary outcomes12
included body weight, waist circumference, sagittal diameter, fasting concentra-
tions of plasma glucose, serum insulin, serum C-peptide plasma GIP, serum tri-
glycerides (TAG), serum total cholesterol, serum high-density lipoprotein (HDL)
cholesterol, serum low-density lipoprotein (LDL) cholesterol, serum alanine-
aminotransferase (ALAT), serum aspartate aminotransferase (ASAT), serum CRP,
serum IL-6, serum TNF-α, whole-blood haemoglobin, white blood cells, whole-
blood lymphocytes, mix of whole-blood monocytes, eosinophils as well as baso-
phils, whole-blood neutrophils, serum IL-6, serum TNF-α, serum zonulin, plasma
citrullin, homoeostatic model assessment for insulin resistance (HOMA-IR),
whole-blood glycated haemoglobin (HbA1c) and targeted serum and faeces
metabolites. In addition, during a standardized meal test measurement of post-
prandial responses of plasma glucose, serum insulin, plasma GLP-2, plasma pep-
tide YY (PYY), plasma free fatty acids (FFA), exhalation of H
2
, untargeted UPLC-
MS and GC-MS urine metabolomics, urine lactulose and mannitol excretion.
Further examinations included measurement of faecal calprotectin, intestinal
transit time, average number of defaecations over the last week, Bristol stool scale
estimates of stool consistence, well-being and gastrointestinal comfort indicators
(bloating), and ex vivo cytokine production in LPS-stimulated whole-blood.
Sample size. Estimations were based on 85% statistical power to detect a difference
of 0.4 standard deviation in metabolic quantitative traits, based on previous
observations from the MetaHit study64. It was estimated that 51 individuals were
needed, but to allow for a 15% dropout after randomization, a total of 60 parti-
cipants were invited for participation. Additionally, based on observed standard
deviations for the MGSs changing during the low-gluten and high-gluten inter-
ventions, we concluded that the number of included subjects was adequate to
provide evidence of a changed intestinal microbiome after a low-gluten diet
compared with a high-gluten diet.
Randomisation. The random allocation sequence was generated by an investigator
without contact to the participants (www.randomization.com). Details of the type
of randomisations and restrictions such as blocking and block size have been
published previously12. The random allocation sequence was implemented by the
dietician using a list of participant IDs matched with allocated sequences.
Blinding. The participants and the investigators involved in outcome assessment
were blinded until the rst examination day. Thereafter, blinding was not possible
due to the nature of the intervention. However, blinding of the allocation sequence
was maintained during sampling of biological materials and initial steps of
bioinformatics and statistical analyses.
Anthropometrics. On the four examination days, before and after each inter-
vention, participants met in the morning after an overnight fast of 10 h and
absenting from physical activity and alcohol consumption for 24 h. In addition,
participants were instructed to avoid smoking and tooth brushing in the morning
of the examination days. Prior to determination of body weight, participants were
asked to empty their bladder and to wear only underwear or light clothing. Body
weight was determined and registered to the nearest 0.05 kg (Lindell Tronic 8000,
Digital Medical Scale, Copenhagen, Denmark). At the rst examination day only,
height was measured with a wall-mounted stadiometer while the participants were
barefooted and it was registered to the nearest 0.1 cm (Hultafors, Sweden). Waist
circumference was measured twice using a exible measuring tape (Meterex,
Lagenfeld, Germany) at the point of the umbilicus after an exhalation and was
registered to the nearest 0.5 cm. Sagittal abdominal diameter was measured twice
using an abdominal calibre (Holtain-Kahn Abdominal Caliper, Crosswell, UK) at
the umbilicus level after an exhalation with participants lying on a at bed with the
legs bent and was registered to the nearest 0.1 cm.
NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x ARTICLE
NATURE COMMUNICATIONS | ( 2018) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Biochemical analyses of fasting blood samples. Blood samples were drawn via
an intravenous cannula in the participantsantecubital vein at all four examination
days. Shortly after collection, the blood samples were stored in ice water, separated
into serum and plasma, and immediately stored at 80 °C until analyses. All blood
sample analyses were performed in one batch at the end of the study to ensure low
variability.
Plasma glucose, whole-blood HbA1c and serum TAG, total-, LDL-, and HDL-
cholesterol, ALAT and ASAT were analysed using automated, enzymatic,
colorimetric assay on ABX Pentra 400 chemistry analyser (ABX Pentra, Horiba
ABX, Montpellier, France). The coefcients of variation (CV) for these analyses
were between 1.3 and 7.2%.
Serum insulin and C-peptide were measured by a chemiluminescent
immunometric assay (Immulite 1000; Siemens Medical Solutions Diagnostics, Los
Angeles, USA). CV was < 5% for both. HOMA-IR was calculated according to
Wallace et al.65 as insulin resistance =glucose in mmol L1× insulin in pmol
L1× 1351. Serum CRP was measured after a 1000× dilution in a high-sensitivity
single-plex assay (MesoScale Discovery®, Gaithersburg, MD, USA) using the
Sector Imager 2400A (MesoScale Discovery®). The lower limit of detection was
4.3 pg mL1. Blood counts of total haemoglobin, leucocytes, neutrophils,
lymphocytes, and others immune cells (including monocytes, mast cells, basophils
and eosinophils) were obtained using a Sysmex KX-21 automated haematology
analyser (Sysmex America Inc., Lincolnshire, Illinois, USA). Serum IL-6 and TNF-
αwere measures by high-sensitivity enzyme-linked immunosorbent assays (ELISA)
(R&D systems, Minneapolis, Minnesota, USA, HSLB00C, HS600B, and HSTA00D,
with detection limits: 0.15 pg mL1and 0.5 pg mL1, respectively). The CV% was
3.6% and 5.2%, respectively.
Plasma citrulline, a marker of enterocyte capacity and mass, was measured
using ultra-performance liquid chromatography tandem mass-spectrometry of
acetonitrile-derived supernatants originally validated and described elsewhere with
a CV% of 2.04.366. Serum zonulin, a marker of tight junction regulation67,was
measured using IDK Zonulin ELISA kit (Immundiagnostik AG, Bensheim,
Germany). The CV% was 7.5%.
Plasma alkylresorcinols, markers of wholegrain wheat, rye and quinoa intake,
were analysed using normal-phase ultra-performance liquid chromatography
tandem mass-spectrometry68.
Dietary intake assessment. Participants completed a 4-day pre-coded dietary
record, developed and used at the National Food Institute at the Technical Uni-
versity of Denmark69,70 to assess the habitual dietary intake in the national dietary
survey. The record was lled out on two weekdays and two weekend days at study
start and at the end of both interventions. Daily intake of total energy, macro-
nutrients, certain food components and food groups were calculated (habitual diet
only without estimates from intake of study products). The gluten content of the
study dietary products was calculated based on data from the food database at the
Danish Food Composition Databank containing 1049 food items71.
Dietary compliance. Participants recorded a study diary, in which they registered
daily consumption (amount and type) of study dietary products throughout both
interventions as well as any deviations from the dietary instructions in the diary. A
trained dietician conducted a follow-up telephone call every second week prior to
home delivery of study dietary products, focusing on consumption of study dietary
products and adherence to the dietary regimens in general. The diary was used as
an objective measure of compliance to the intervention and to calculate absolute
consumption of study dietary products. In addition, the concentration of alkylre-
sorcinols in the blood was analysed as a measure of compliance, since these are
biomarkers of grains13. The study diary was also used for noting any illness and use
of antibiotics, during the interventions.
Dietary bre composition of the two diets. The dietary bre composition of the
two diets were determined by measuring the monosaccharide composition of a
representative meal of each distinct diet, the resistant starch composition of the
dietary study products, and the FODMAP (fermentable oligosaccharides, dis-
accharides, monosaccharides and polyols) composition of the provided low-gluten
and the high-gluten study products. The details are available in Supplementary
Methods.
Faecal sample collection and DNA extraction. Faecal samples were collected in
the morning of the four examination days and immediately stored at 5 °C for a
maximum of 24 h before equal volume of sterile water was added and the sample
was homogenised. The homogenised sample was aliquoted to cryotubes, and stored
at 80 °C. Microbial DNA was extracted from the faecal samples as previously
reported72.
Metagenomic sequencing and quantitative PCR. The community DNA from all
faecal samples was sequenced by metagenomics sequencing. In addition, quanti-
cation of Bidobacterium spp. and total bacterial load in all faecal samples was
performed by quantitative PCR. Details are available in Supplementary Methods.
Ex vivo cytokine production after stimulation with LPS. Within 30 min of blood
sampling, 50 μL of whole-blood was LPS stimulated in triplicates after having been
diluted 1:10 in RPMI medium (LONZA, BE12-167F) supplemented with LPS
(Sigma-Aldrich, L2645-1MG) in a nal concentration of 1 μgmL
1. Samples were
incubated for ~24 h at 37 °C and 5% CO
2
in order to determine ex vivo cytokine
production. After incubation supernatants were harvested and stored at 80 °C
until ELISA measurements of IL-1β, IL-6, TNF-αand IFN-γ(R&D Systems,
DY201, DY206, and DY210, respectively)73. The CV% was 3.65.2%.
Standardised meal test. On the four examination days, participants were lying
and resting for at least 10 min before blood samples were drawn in the at least 8 h
fasting state (t=0 min) and postprandial (t=30, 60, 120 and 180 min) after
consumption of the same standardised breakfast, no matter which intervention the
study participant was allocated to. The meal consisted of white wheat bread, a
pastry, butter, jam, cheese and 200 mL water (3000 kJ, 52.6 E% fat, 39.7 E%
carbohydrate, 7.8 E% protein) and a standardised drink contai ning lactulose (5 g)
and mannitol (2 g). Participants rated their well-being twice at fasting, and every
30 min following the standardised breakfast using a 100 mm visual analogue scale
(VAS) with the most positive and the most negative ratings at each end of the line.
Biochemical analyses of postprandial blood samples. Upon the standardised
meal test, plasma glucose and serum insulin were measured in all postprandial
blood samples as described above and plasma PYY, plasma GIP and plasma GLP-2
were measured in all postprandial blood samples as specied in Supplementary
Methods.
Exhalation of hydrogen. Hydrogen exhalation was measured twice at fasting, and
every 30 min following the standardised breakfast and drink (t=30, 60, 90, 120,
150 and 180 min). Breath hydrogen was measured in exhaled breath as a proxy
measure of colonic fermentation using a handheld calibrated Gastro+Gastrolyzer®
(Bedfont Scientic Ltd.). Participants were instructed to breath in deeply; hold their
breath for 15 s and then exhale at a steady pace into the cardboard mouthpiece of
the device until their lungs felt empty.
Visual analogue scoring of gastrointestinal indicators. Participants rated their
well-being and gastrointestinal symptoms (bloating) during the past week using
visual analogue scoring. The reliability and validity have been examined and a VAS
score is considered to be a methodologically reliable measure of gastrointestinal
comfort/discomfort74. Furthermore, participants provided information on smok-
ing, intake of medications and dietary supplements and assessed their stool con-
sistence on a 7-point scale (Bristol stool form scale) as well as their defaecation
frequency75.
Intestinal transit time. For six consecutive days before examination days 1, 2 and
4, the participants ingested 24 non-absorbable radio-opaque transit plastic ring
markers in the morning on a daily basis to ensure saturation and lled in a
defecation diary. Abdominal radiographs were performed in the afternoon at
Frederiksberg Hospital, Copenhagen, Denmark on day 7 (examination day: the
same day when faecal, blood and urine samples were collected), 30 h after the last
transit marker intake. Intestinal transit time was estimated (ranked) based on the
number of visible plastic markers on the obtained abdominal X-ray, adjusted for
time since last marker ingestion. This was calculated as follows: number of markers
counted from the X-ray lm × the number of h between last marker ingest ion and
radiograph divided by 24 (daily dose of markers). The resulting relative transit time
estimates enabled us to rank the participants according to transit time as
reported22.
Faecal calprotectin. Faecal calprotectin is used as a marker of inammation of the
small intestine, large bowel or the stomach. Calprotectin content in stools was
measured using CALPROLABTM Calprotectin ELISA (ALP) (Calpro AS, Oslo,
Norway), which is an ELISA based on polyclonal antibodies to human calprotectin
with a reported CV% of 6.18.7% (S100A8/A9).
Gut permeability assessment. After ingestion of the standardised breakfast and
drink containing lactulose (5 g) and mannitol (2 g), urine was collected for 4 h.
During the time of collection, urine was stored in the fridge. After collection, the
urine samples were aliquoted into 2.0 mL tubes and stored at 80 °C. Quanti-
cation of lactulose in urine samples was performed by chromatographic analysis.
Briey, high-performance anion-exchange chromatography was performed with a
Dionex CarboPac MA1 BioLC Analytical 4×250 mm column. The carbohydrate
separation was performed using a Dionex CarboPac MA1 BioLC Guard 4×50 mm
column (Dionex Corp, Sunnyvale, CA, USA). The samples were eluted with 50 mM
NaOH at a ow rate of 1 mL min1. Urinary excretion of mannitol was quantied
using spectrophotometric analysis on a ABX Pentra 400 (Horiba Medical, Cali-
fornia, USA). The CV% was 12.8 for lactulose and 0.8 for mannitol. The percentage
of excreted lactulose and mannitol in urine after administration of the liquid
formulation was evaluated, and the lactulose/mannitol ratio was calculated for each
sample.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x
10 NATURE COMMUNICATIONS | (201 8) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Collection of urine samples. Upon arrival on each examination day, participants
emptied their bladder. Urine was collected for 4 h after the standard ised breakfast
(containing approximately 3000 kJ, 52.6 E% fat, 39.7 E% carbohydrates, and 78E
% protein) and the lactulose-mannitol containing drink. Urine was stored in the
fridge during collection, pooled, mixed and aliquoted into 2.0 mL tubes and stored
at 80 °C. A complete set of the 4 h urine samples was available from 51 of the
completing participants.
Urine creatinine measurement. Creatinine concentrations were measured using
urinary creatinine ELISA kit from Arbor Assays (Ann Arbor, Michigan, USA). All
samples were diluted 1:20 and measured in duplicates (CV% was 1.7%). The range
of the creatinine standard curve was 0.3120 mg dL1. Creatinine concentrations
were used to adjust the injection volume of each urine sample when analysed by
UPLC-MS as well as to normalise GC-MS data to account for the dilution of urine.
Metabolomics. Untargeted urine metabolomics as well as targeted serum and
faeces metabolomics quantifying short-chain fatty acids, fatty acids, acyl-carnitines,
BAIBA and metabolites associated with the microbiota-gut-brain axis including
serotonin, kynurenine, glutamate, γ-aminobutyric acid were performed by UPLC-
MS and GC-MS. Details are available in Supplementary Methods.
Statistical analyses. All statistical analyses were performed in R version 3.1 (The
R Foundation for Statistical Computing, 2012, Vienna, Austria)76. Available-case
analyses were carried out for all outcomes. The effects of the interventions on all
outcomes were analysed by a linear mixed model (LMM) using the lme4 R-
package77 with participant-specic and within-period participant-specic random
effects. The model included an interventionvisit interaction and adjustment for
age and gender as xed effects. In addition, adjustment for intestinal transit time
was included since this parameter recently has been reported to be an important
confounder22. The effects of the intervention were assessed using the multcomp R-
package78. Four individuals underwent antibiotics treatment during the trial and
visits following antibiotics treatment were excluded from all statistics. Further
details on statistical analyses are available in Supplementary Methods.
Data availability
The raw Illumina read data for all samples have been deposited in the Short Read
Archive under the Bioproject: PRJNA491335. Other data supporting the ndings of
the study are available in this article and its Supplementary Information les, or
from the corresponding authors upon request.
Received: 10 December 2017 Accepted: 5 October 2018
References
1. Kagnoff, M. F. Celiac disease: pathogenesis of a model immunogenetic disease.
J. Clin. Invest. 117,4149 (2007).
2. Wieser, H. The precipitating factor in coeliac disease. Baillieres Clin.
Gastroenterol. 9, 191207 (1995).
3. Zhang, L. et al. Effects of gliadin consumption on the intestinal microbiota and
metabolic homeostasis in mice fed a high-fat diet. Sci. Rep. 7, 44613 (2017).
4. Juhász, A. et al. Genome mapping of seed-borne allergens and
immunoresponsive proteins in wheat. Sci. Adv. 4, eaar8602 (2018).
5. Vazquez-Roque, M. I. et al. A controlled trial of gluten-free diet in patients
with irritable bowel syndrome-diarrhea: effects on bowel frequency and
intestinal function. Gastroenterology 144, 903911.e3 (2013).
6. Orlando, A., Linsalata, M., Notarnicola, M., Tutino, V. & Russo, F.
Lactobacillus GG restoration of the gliadin induced epithelial barrier
disruption: the role of cellular polyamines. Bmc. Microbiol. 14, 19 (2014).
7. Jin, M., Xu, C. & Zhang, X. The effect of tryptophol on the bacteriophage
infection in high-temperature environment. Appl. Microbiol. Biotechnol. 99,
81018111 (2015).
8. Caminero, A. et al. Diversity of the cultivable human gut microbiome involved
in gluten metabolism: isolation of microorganisms with potential interest for
coeliac disease. FEMS Microbiol. Ecol. 88, 309319 (2014).
9. Caminero, A. et al. A gluten metabolism study in healthy individuals shows
the presence of faecal glutenasic activity. Eur. J. Nutr. 51, 293299 (2012).
10. De Palma, G., Nadal, I., Collado, M. C. & Sanz, Y. Effects of a gluten-free diet
on gut microbiota and immune function in healthy adult human subjects. Br.
J. Nutr. 102, 1154 (2009).
11. Bonder, M. J. et al. The inuence of a short-term gluten-free diet on the
human gut microbiome. Genome Med. 8, 45 (2016).
12. Ibrügger, S. et al. Two randomized cross-over trials assessing the impact of
dietary gluten or wholegrain on the gut microbiome and host metabolic
health. J. Clin. Trials 4, 178 (2014).
13. Lind, M. V. et al. Plasma alkylresorcinols reect gluten intake and distinguish
between gluten-rich and gluten-poor diets in a population at risk of metabolic
syndrome. J. Nutr. 146, 19911998 (2016).
14. Hoppe, C. et al. Intake and sources of gluten in 20- to 75-year-old Danish
adults: a national dietary survey. Eur. J. Nutr. 56, 107117 (2017).
15. Biesiekierski, J. R. et al. No effects of gluten in patients with self-reported non-
celiac gluten sensitivity after dietary reduction of fermentable, poorly
absorbed, short-chain carbohydrates. Gastroenterology 145, 320328.e3
(2013).
16. Li, J. et al. An integrated catalog of reference genes in the human gut
microbiome. Nat. Biotechnol. 32, 834841 (2014).
17. Nielsen, H. B. et al. Identication and assembly of genomes and genetic
elements in complex metagenomic samples without using reference genomes.
Nat. Biotechnol. 32, 822828 (2014).
18. Kanehisa, M., Sato, Y., Kawashima, M., Furumichi, M. & Tanabe, M. KEGG as
a reference resource for gene and protein annotation. Nucleic Acids Res. 44,
D457D462 (2016).
19. Kanehisa, M. et al. Data, information, knowledge and principle: back to
metabolism in KEGG. Nucleic Acids Res. 42, D199D205 (2014).
20. Vieira-Silva, S. et al. Speciesfunction relationships shape ecological properties
of the human gut microbiome. Nat. Microbiol. 1, 16088 (2016).
21. Nakamura, N., Lin, H. C., McSweeney, C. S., Mackie, R. I. & Gaskins, H. R.
Mechanisms of microbial hydrogen disposal in the human colon and
implications for health and disease. Annu. Rev. Food Sci. Technol. 1, 363395
(2010).
22. Roager, H. M. et al. Colonic transit time is related to bacterial metabolism and
mucosal turnover in the gut. Nat. Microbiol. 1, 16093 (2016).
23. Karra, E., Chandarana, K. & Batterham, R. L. The role of peptide YY in
appetite regulation and obesity. J. Physiol. 587,1925 (2009).
24. Matikainen, N. et al. Minor contribution of endogenous GLP-1 and GLP-2 to
postprandial lipemia in obese men. PLoS. ONE. 11, e0145890 (2016).
25. Canfora, E. E. et al. Colonic infusions of short-chain fatty acid mixtures
promote energy metabolism in overweight/obese men: a randomized
crossover trial. Sci. Rep. 7, 2360 (2017).
26. Guo, Y. et al. Physiological evidence for the involvement of peptide YY in the
regulation of energy homeostasis in humans*. Obesity 14, 15621570 (2006).
27. Boey, D. et al. PYY transgenic mice are protected against diet-induced and
genetic obesity. Neuropeptides 42,1930 (2008).
28. Roberts, L. D. et al. β-Aminoisobutyric acid induces browning of white fat and
hepatic β-oxidation and is inversely correlated with cardiometabolic risk
factors. Cell. Metab. 19,96108 (2014).
29. Kennedy, P. J., Cryan, J. F., Dinan, T. G. & Clarke, G. Kynurenine pathway
metabolism and the microbiota-gut-brain axis. Neuropharmacology 112,
399412 (2017).
30. Mazzoli, R. & Pessione, E. The neuro-endocrinological role of microbial
glutamate and GABA signaling. Front. Microbiol. 7, 1934 (2016).
31. van Hees, N. J. M. et al. Essential amino acids in the gluten-free diet and
serum in relation to depression in patients with celiac disease. PLoS. ONE. 10,
e0122619 (2015).
32. Agudelo, L. Z. et al. Kynurenic acid and Gpr35 regulate adipose tissue energy
homeostasis and inammation. Cell. Metab. 27, 378392.e5 (2018).
33. Wang, J. et al. Kynurenic acid as a ligand for orphan G protein-coupled
receptor GPR35. J. Biol. Chem. 281, 2202122028 (2006).
34. Collado, M. C., Donat, E., Ribes-Koninckx, C., Calabuig, M. & Sanz, Y.
Imbalances in faecal and duodenal Bidobacterium species composition in
active and non-active coeliac disease. Bmc. Microbiol. 8, 232 (2008).
35. Di Cagno, R. et al. Different fecal microbiotas and volatile organic compounds
in treated and untreated children with celiac disease. Appl. Environ. Microbiol.
75, 39633971 (2009).
36. Staudacher, H. M. et al. A diet low in FODMAPs reduces symptoms in
patients with irritable bowel syndrome and a probiotic restores
bidobacterium species: a randomized controlled trial. Gastroenterology 153,
936947 (2017).
37. McIntosh, K. et al. FODMAPs alter symptoms and the metabolome of patients
with IBS: a randomised controlled trial. Gut 66, 12411251 (2017).
38. Bennet, S. M. P. et al. Multivariate modelling of faecal bacterial proles of
patients with IBS predicts responsiveness to a diet low in FODMAPs. Gut 67,
872881 (2018).
39. Smits, S. A. et al. Seasonal cycling in the gut microbiome of the Hadza hunter-
gatherers of Tanzania. Science 357, 802806 (2017).
40. De Filippo, C. et al. Impact of diet in shaping gut microbiota revealed by a
comparative study in children from Europe and rural Africa. Proc. Natl Acad.
Sci. USA 107, 1469114696 (2010).
41. Belenguer, A. et al. Two routes of metabolic cross-feeding between
Bidobacterium adolescentis and butyrate-producing anaerobes from the
human gut. Appl. Environ. Microbiol. 72, 35933599 (2006).
42. Falony, G., Vlachou, A., Verbrugghe, K. & De Vuyst, L. Cross-feeding between
Bidobacterium longum BB536 and acetate-converting, butyrate-producing
NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x ARTICLE
NATURE COMMUNICATIONS | ( 2018) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications 11
Content courtesy of Springer Nature, terms of use apply. Rights reserved
colon bacteria during growth on oligofructose. Appl. Environ. Microbiol. 72,
78357841 (2006).
43. Rivière, A., Gagnon, M., Weckx, S., Roy, D. & De Vuyst, L. Mutual cross-
feeding interactions between Bidobacterium longum subsp. longum
NCC2705 and eubacterium rectale atcc 33656 explain the bidogenic and
butyrogenic effects of arabinoxylan oligosaccharides. Appl. Environ. Microbiol.
81, 77677781 (2015).
44. Chassard, C. & Bernalier-Donadille, A. H
2
and acetate transfers during xylan
fermentation between a butyrate-producing xylanolytic species and
hydrogenotrophic microorganisms from the human gut. FEMS Microbiol.
Lett. 254, 116122 (2006).
45. Rey, F. E. et al. Dissecting the in vivo metabolic potential of two human gut
acetogens. J. Biol. Chem. 285, 2208222090 (2010).
46. Lawson, P. A., Blaut, M., Simmering, R., Collins, M. D. & Taras, D.
Reclassication of Eubacterium formicigenerans Holdeman and Moore 1974
as Dorea formicigenerans gen. nov., comb. nov., and description of Dorea
longicatena sp. nov., isolated from human faeces. Int. J. Syst. Evol. Microbiol.
52, 423428 (2002).
47. Van den Abbeele, P. et al. Arabinoxylans and inulin differentially modulate
the mucosal and luminal gut microbiota and mucin-degradation in
humanized rats. Environ. Microbiol. 13, 26672680 (2011).
48. Neyrinck, A. M. et al. Prebiotic effects of wheat arabinoxylan related to the
increase in bidobacteria, Roseburia and Bacteroides/Prevotella in diet-
induced obese mice. PLoS. ONE. 6, e20944 (2011).
49. Hughes, S. A. et al. In vitro fermentation by human fecal microora of wheat
arabinoxylans. J. Agric. Food Chem. 55, 45894595 (2007).
50. Sanchez, J. I. et al. Arabinoxylan-oligosaccharides (AXOS) affect the protein/
carbohydrate fermentation balance and microbial population dynamics of the
Simulator of Human Intestinal Microbial Ecosystem. Microb. Biotechnol. 2,
101113 (2009).
51. Damen, B. et al. Prebiotic effects and intestinal fermentation of cereal
arabinoxylans and arabinoxylan oligosaccharides in rats depend strongly on
their structural properties and joint presence. Mol. Nutr. Food Res. 55,
18621874 (2011).
52. Dervilly-Pinel, G., Rimsten, L., Saulnier, L., Andersson, R. & Åman, P. Water-
extractable arabinoxylan from pearled ours of wheat, barley, rye and triticale.
Evidence for the presence of ferulic acid dimers and their involvement in gel
formation. J. Cereal Sci. 34, 207214 (2001).
53. Corazza, G. R., Strocchi, A. & Gasbarrini, G. Fasting breath hydrogen in celiac
disease. Gastroenterology 93,5358 (1987).
54. Soares, F. L. P. et al. Gluten-free diet reduces adiposity, inammation and
insulin resistance associated with the induction of PPAR-alpha and PPAR-
gamma expression. J. Nutr. Biochem. 24, 11051111 (2013).
55. Freire, R. H. et al. Wheat gluten intake increases weight gain and adiposity
associated with reduced thermogenesis and energy expenditure in an animal
model of obesity. Int. J. Obes. (Lond.). 40, 479486 (2016).
56. Haupt-Jorgensen, M., Buschard, K., Hansen, A. K., Josefsen, K. & Antvorskov,
J. C. Gluten-free diet increases beta-cell volume and improves glucose
tolerance in an animal model of type 2 diabetes. Diabetes Metab. Res. Rev. 32,
675684 (2016).
57. Rune, I. et al. Modulating the gut microbiota improves glucose tolerance,
lipoprotein prole and atherosclerotic plaque development in ApoE-decient
mice. PLoS. ONE. 11, e0146439 (2016).
58. Latz, E., Xiao, T. S. & Stutz, A. Activation and regulation of the
inammasomes. Nat. Rev. Immunol. 13, 397411 (2013).
59. Chen, G. Y. Role of Nlrp6 and Nlrp12 in the maintenance of intestinal
homeostasis. Eur. J. Immunol. 44, 321327 (2014).
60. Elinav, E. et al. NLRP6 inammasome regulates colonic microbial ecology and
risk for colitis. Cell 145, 745757 (2011).
61. Biagi, F. & Corazza, G. R. Mortality in celiac disease. Nat. Rev. Gastroenterol.
Hepatol. 7, 158162 (2010).
62. Lebwohl, B. et al. Long term gluten consumption in adults without celiac
disease and risk of coronary heart disease: prospective cohort study. BMJ 357
j1892, https://doi.org/10.1136/bmj.j1892 (2017).
63. Aadahl, M., Zacho, M., Linneberg, A., Thuesen, B. H. & Jorgensen, T.
Comparison of the Danish step test and the watt-max test for estimation of
maximal oxygen uptake: the Health2008 study. Eur. J. Prev. Cardiol. 20,
10881094 (2013).
64. Le Chatelier, E. et al. Richness of human gut microbiome correlates with
metabolic markers. Nature 500, 541546 (2013).
65. Wallace, T. M., Levy, J. C. & Matthews, D. R. Use and abuse of HOMA
modeling. Diabetes Care 27, 14871495 (2004).
66. Demacker, P. N. et al. Plasma citrulline measurement using UPLC
tandem mass-spectrometry to determine small intestinal enterocyte
pathology. J. Chromatogr. B. Analyt. Technol. Biomed. Life. Sci. 877,
387392 (2009).
67. Qi, Y. et al. Intestinal permeability biomarker zonulin is elevated in healthy
aging. J. Am. Med. Dir. Assoc. 18, 810.e1810.e4 (2017).
68. Ross, A. B. et al. A high-throughput method for liquid
chromatographytandem mass spectrometry determination of plasma
alkylresorcinols, biomarkers of whole grain wheat and rye intake. Anal.
Biochem. 499,17 (2016).
69. Rasmussen, L. B., Matthiessen, J., Biltoft-Jensen, A. & Tetens, I. Characteristics
of misreporters of dietary intake and physical activity. Public Heal. Nutr. 10,
230237 (2007).
70. Biltoft-Jensen, A. et al. Validation of the Danish 7-day pre-coded food diary
among adults: energy intake v. energy expenditure and recording length. Br. J.
Nutr. 102, 18381846 (2009).
71. Saxholt, E. et al. Danish Food Composition Databank, revision 7. Department
of Nutrition, National Food Institute, Technical University of Denmark (2008).
72. Godon, J. J., Zumstein, E., Dabert, P., Habouzit, F. & Moletta, R. Molecular
microbial diversity of an anaerobic digestor as determined by small-subunit
rDNA sequence analysis. Appl. Environ. Microbiol. 63, 28022813 (1997).
73. Damsgaard, C. T., Lauritzen, L., Calder, P. C., Kjær, T. M. R. & Frøkiær, H.
Whole-blood culture is a valid low-cost method to measure monocytic
cytokinesA comparison of cytokine production in cultures of human whole-
blood, mononuclear cells and monocytes. J. Immunol. Methods 340,95101
(2009).
74. Gallagher, E. J., Bijur, P. E., Latimer, C. & Silver, W. Reliability and validity of
a visual analog scale for acute abdominal pain in the ED. Am. J. Emerg. Med.
20, 287290 (2002).
75. Lewis, S. J. & Heaton, K. W. Stool form scale as a useful guide to intestinal
transit time. Scand. J. Gastroenterol. 32, 920924 (1997).
76. R. Core Team. R: A language and environment for statistical computing. R
Foundation for Statistical Computing. 3, 201 (2013).
77. Bates, D., Maechler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects
models using lme4. J. Stat. Softw. 67,148 (2015).
78. Hothorn, T., Bretz, F. & Westfall, P. Simultaneous inference in general
parametric models. Biom. J. 50, 346363 (2008).
Acknowledgements
The authors are indebted to Jeffrey Edward Skiby for administrative and secretarial
support, Annemette Forman, Tina H. Lorentzen, Sarah Ben Soltane, Josue Leonardo
Castro-Mejia, Charlotte Holm Brodersen, Pernille Lærke Bjørndal Hollænder, Anne
Marie Raabyemagle, Kate V. Vibefelt, Neslihan Bicen, Morgan Han and Elizabeth
McKenzie for excellent biotechnical support, as well as Axel Kornerup and Lasse Ingvar
Hellgren who participated in pertinent academic discussions related to this trial as
members of the scientic committee. Sequencing was carried out at the Technical
University of Denmark in-house facility (DTU Multi-Assay Core, DMAC), in colla-
boration with BGI Copenhagen. We also gratefully acknowledge the Danish National
Supercomputer for Life Sciences Computerome (computerome.dtu.dk) for the com-
putational resources to perform the sequence analyses and storage. Selected dietary
products were sponsored by Kohberg, Lantmännen, AXA, Wasa, Urtekram, Finax, and
Doves Farm. The study was supported by the Innovation Fund Denmark (grant no. 11-
116163/0603-00487B; Center for Gut, Grain and Greens (3G Center)) and The Novo
Nordisk Foundation. The Novo Nordisk Foundation Center for Basic Metabolic
Research is an independent research centre at the University of Copenhagen and is partly
funded by an unrestricted donation from the Novo Nordisk Foundation. The study
sponsors and the employers had no inuence on the design, applied methods, data
generation and analysis, or in the decision to publish.
Author contributions
O.P. conceived the concept of trial while the conductance of trial was supervised by O.P.
and L.L. T.H. contributed to trial design. The daily management responsible of the trial
running was R.J.G., O.P., M.K and S.I. H.F., S.B., A.B.R., M.C.L., A.F.C., J.H.P., M.H.S., J.
J.H., B.H., R.B.M., M.V.L., S.I., M.K. and R.G. were involved in obtaining biochemical
and physiological measures. The dietary and nutritional analyses were done by D.S., C.
H., A.S.M., A.B., J.H. and I.T. while M.D.D., V.C., J.M.L. and M.I.B. were involved in the
generation of microbiota data. S.V-B., H.L.F. and H.M.R. monitored and analysed the
urine metabolome. J.H. and N.J.F. analysed the serum and faecal metabolome. K.H.A., T.
N. and M.L.M. undertook host phenotype quality control analyses. C.R. and C.E. devised
the statistical models. L.B.S.H., G.F., M.V.C. and S.V.S. did the bioinformatics analyses
supervised by R.G., J.R. and H.B.N. The statistical analyses of host physiology and
nutritional data were performed by L.B.S.H. and N.B.S. Expert supervision was per-
formed by J.J.R., B.H., J.J.H., A.L., K.B., T.S-B, H.V. and K.K. The core analytical and
writing team consisting of O.P., T.R.L., R.G., H.B.N., M.I.B., N.B.S., H.M.R. and L.B.S.H.
led the data compiling and the interpretation of trial outcome. L.B.S.H., H.M.R. and N.B.
S. undertook the integrative data analyses and drafted the manuscript with substantial
contributions from O.P., T.R.L., L.L., H.F. and S.B. All authors contributed to and
approved the nal manuscript.
Additional information
Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467-
018-07019-x.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x
12 NATURE COMMUNICATIONS | (201 8) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Competing interests: The authors declare no competing interests.
Reprints and permission information is available online at http://npg.nature.com/
reprintsandpermissions/
Publishers note: Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional afliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the articles Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
articles Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2018
Lea B.S. Hansen
1
, Henrik M. Roager
2,3
, Nadja B. Søndertoft
4
, Rikke J. Gøbel
4
, Mette Kristensen
3
,
Mireia Vallès-Colomer
5,6
, Sara Vieira-Silva
5,6
, Sabine Ibrügger
3
, Mads V. Lind
3
, Rasmus B. Mærkedahl
3,7
,
Martin I. Bahl
2
, Mia L. Madsen
4
, Jesper Havelund
8
, Gwen Falony
5,6
, Inge Tetens
3
, Trine Nielsen
4
,
Kristine H. Allin
4
, Henrik L. Frandsen
2
, Bolette Hartmann
9
, Jens Juul Holst
4
, Morten H. Sparholt
10
, Jesper Holck
11
,
Andreas Blennow
12
, Janne Marie Moll
13
, Anne S. Meyer
11
, Camilla Hoppe
2
, Jørgen H. Poulsen
14
,
Vera Carvalho
2
, Domenico Sagnelli
12
, Marlene D. Dalgaard
13
, Anders F. Christensen
10
,
Magnus Christian Lydolph
15
, Alastair B. Ross
16
, Silas Villas-Bôas
17
, Susanne Brix
13
, Thomas Sicheritz-Pontén
1
,
Karsten Buschard
18
, Allan Linneberg
19
, Jüri J. Rumessen
20
, Claus T. Ekstrøm
21
, Christian Ritz
3
,
Karsten Kristiansen
22
, H. Bjørn Nielsen
23
, Henrik Vestergaard
4
, Nils J. Færgeman
8
, Jeroen Raes
5,6
,
Hanne Frøkiær
7
, Torben Hansen
4
, Lotte Lauritzen
3
, Ramneek Gupta
1
, Tine Rask Licht
2
& Oluf Pedersen
4
1
Department of Bio and Health Informatics, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark.
2
National Food Institute, Technical
University of Denmark, DK-2800 Kgs. Lyngby, Denmark.
3
Department of Nutrition, Exercise and Sports, Faculty of Science, University of
Copenhagen, DK-1958 Frederiksberg, Denmark.
4
The Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen,
DK-2200 Copenhagen, Denmark.
5
Department of Microbiology and Immunology, KU LeuvenUniversity of Leuven, Rega Institute, 3000 Leuven,
Belgium.
6
VIB, Center for Microbiology, 3000 Leuven, Belgium.
7
Department of Veterinary Disease Biology, Faculty of Science, University of
Copenhagen, DK-1958 Frederiksberg, Denmark.
8
Department of Biochemistry and Molecular Biology, University of Southern Denmark, DK-5230
Odense, Denmark.
9
Department of Biomedical Sciences, University of Copenhagen, Copenhagen DK-2200, Denmark.
10
Department of Radiology,
Bispebjerg Hospital, Copenhagen DK-2400, Denmark.
11
Department of Chemical and Biochemical Engineering, Technical University of Denmark,
DK-2800 Kgs. Lyngby, Denmark.
12
Department of Plant and Environmental Sciences, University of Copenhagen, DK-1958 Frederiksberg, Denmark.
13
Department of Biotechnology and Biomedicine, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark.
14
Department of Clinical
Biochemistry, Copenhagen University Hospital Hvidovre, DK-2650 Hvidovre, Denmark.
15
Department of Autoimmunology & Biomarkers, Statens
Serum Institut, DK-2300 Copenhagen, Denmark.
16
Department of Biology and Biological Engineering, Chalmers University of Technology, 412 96
Gothenburg, Sweden.
17
School of Biological Sciences, The University of Auckland, 1010 Auckland, New Zealand.
18
Bartholin Institute, Rigshospitalet,
DK-2200 Copenhagen, Denmark.
19
Research Centre for Prevention and Health, The Capital Region of Denmark, DK-2000 Frederiksberg, Denmark.
20
Research Unit and Department of Gastroenterology, Herlev and Gentofte Hospital, the Capital Region of Denmark, 2730 Herlev, Denmark.
21
Biostatistics, Department of Public Health, University of Copenhagen, DK-1014 Copenhagen, Denmark.
22
Laboratory of Genomics and Molecular
Biomedicine, Department of Biology, University of Copenhagen, DK-2100 Copenhagen, Denmark.
23
Clinical-Microbiomics A/S, DK-2200
Copenhagen, Denmark. These authors contributed equally: Lea B.S. Hansen, Henrik M. Roager, Nadja B. Søndertoft, Rikke J. Gøbel.
NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-07019-x ARTICLE
NATURE COMMUNICATIONS | ( 2018) 9:4630 | DOI: 10.1038/s41467-018-07019-x | www.nature.com/naturecommunications 13
Content courtesy of Springer Nature, terms of use apply. Rights reserved
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
... However, recent diagnostic developments have surfaced the gluten-mediated allergenic responses, which ultimately cause celiac disease (CD) and immunoglobulin E-mediated Wheat Allergy (IWA) in humans. The compositional combination of selective peptides (glutamin and proline) and glycoproteins (gliadin) render gluten partially resistant to proteolytic enzymes present in human gastrointestinal tract (Hansen et al. 2018). Unattended, these moieties may interact in the small intestine with immune cells (Herrán et al. 2017), permeability proteins (Vazquez-Roque et al. 2013), and microbial metabolism (Caminero et al. 2014). ...
... The recent human trials describing prophylactic/therapeutic effects of GFD on otherwise-allergic patients have been enlisted in Table 3. The best therapeutic intervention for reducing gluten allergies is the strict and life-long consumption of GFD, as it resolves the local (intestinal), systemic (extra-intestinal symptoms, negativity of autoantibodies and serum biomarkers) and physiological alterations associated with gluten allergies (Aziz et al. 2016;Zanwar et al. 2016;Hansen et al. 2018;Daveson et al. 2020;Kaur et al. 2020;Kosendiak et al. 2020;Neuman et al. 2020;Pinto-Sanchez et al. 2021). However, some probable disadvantages (like negative impacts on the psychological health along-with frequent constipation) may also be seen with few tested individuals (Caio et al. 2019). ...
... Hence, it becomes utmost important to study the interactions among GFD and human microbiota (intestinal or extra-intestinal), besides other genetic, biochemical and physiological parameters, while testing the therapeutic interventions of such diets on gluten-related disorders. On the same line, researchers in a recent Danish human trial (Hansen et al. 2018) studied the microbiome-modulatory role of GFD and the positive health benefit associated with the therapy. ...
Article
Gluten-enteropathy affects a significant number of people, making gluten a major concern in the food industry. With medical advancements, the diagnosis of allergies is becoming easier, and people who are allergic to gluten are recommended a complete gluten-free diet. Since wheat provides a major part of the energy and nutrition in the diet, its elimination affects nutrition intake of allergic population. Food scientists are working to formulate products using protein-rich gluten-free grains with quality attributes at par with gluten-containing products. Focused research has been done to provide nutrition and a variety of food to people suffering from gluten-related disorders. Efforts are being made to remove the gluten from the wheat and other gluten-containing grains, while applying different processing/treatments to enhance the properties of gluten-free grains. Hence, the present review summarizes the importance, processing, and products of different gluten-free grains. It also highlights the digestibility of gluten-free grains with clinical trials and gluten elimination strategies for gluten-containing grains.
... Reference-based computational approaches complement assembly by relying on annotated reference sequence information to accurately identify and quantify the known taxa and genes present in a microbiome by homology instead [4][5][6][7] . This set of methods enabled deep exploration of human microbiomes and the discovery of microbial associations with multiple health conditions [10][11][12][13][14][15][16][17][18] and dietary patterns [19][20][21][22][23] , as well as the characterization of the evolution and transmission of microbial species and strains [24][25][26][27][28][29] . However, reference-based methods can only detect well-characterized and cataloged microbial species included in available reference databases, which typically only represent a fraction of the community members across environments, thus limiting the interpretation of shotgun metagenomes 30 . ...
... We used MetaPhlAn 4 to extend links between the gut microbiome, diet, and host metabolism [19][20][21][22][23]62 by re-analyzing metagenomes from 1,001 deeply phenotyped individuals in the ZOE PREDICT 1 study 22 . As in the original study, strengths of association between the microbiome and both dietary and cardiometabolic host variables were evaluated by testing the predictive power of random forest (RF) classifiers and regressors trained on the taxonomic profiles (see Methods). ...
Preprint
Full-text available
Metagenomic assembly enables novel organism discovery from microbial communities, but from most metagenomes it can only capture few abundant organisms. Here, we present a method - MetaPhlAn 4 - to integrate information from both metagenome assemblies and microbial isolate genomes for improved and more comprehensive metagenomic taxonomic profiling. From a curated collection of 1.01M prokaryotic reference and metagenome-assembled genomes, we defined unique marker genes for 26,970 species-level genome bins, 4,992 of them taxonomically unidentified at the species level. MetaPhlAn 4 explains ~20% more reads in most international human gut microbiomes and >40% in less-characterized environments such as the rumen microbiome, and proved more accurate than available alternatives on synthetic evaluations while also reliably quantifying organisms with no cultured isolates. Application of the method to >24,500 metagenomes highlighted previously undetected species to be strong biomarkers for host conditions and lifestyles in human and mice microbiomes, and showed that even previously uncharacterized species can be genetically profiled at the resolution of single microbial strains. MetaPhlAn 4 thus integrates the novelty of metagenomic assemblies with the sensitivity and fidelity of reference-based analyses, providing efficient metagenomic profiling of uncharacterized species and enabling deeper and more comprehensive microbiome biomarker detection.
... There are many reports on the relationship between food and gut microbiome changes or the association between microbiome composition changes and disease. The MetaHIT Consortium extracted DNA from human feces and compared a group with a low number of gut microbiome genes and a group with a high number of gut microbiome genes to calculate a low gut microbiome gene [43]. In the treated group, high inflammatory cytokine secretion and high obesity rates were confirmed. ...
Article
Full-text available
Obesity associated with a Western diet such as a high-fat diet (HFD) is a known risk factor for inflammatory bowel disease (IBD) and colorectal cancer (CRC). In this study, we aimed to develop fecal microbiome data-based deep learning algorithms for the risk assessment of colorectal diseases. The effects of a HFD and a candidate food (Nypa fruticans, NF) on IBD and CRC risk reduction were also evaluated. Fecal microbiome data were obtained from 109 IBD patients, 111 CRC patients, and 395 healthy control (HC) subjects by 16S rDNA amplicon sequencing. IBD and CRC risk assessment prediction models were then constructed by deep learning algorithms. Dietary effects were evaluated based on fecal microbiome data from rats fed on a regular chow diet (RCD), HFD, and HFD plus ethanol extracts or water extracts of NF. There were significant differences in taxa when IBD and CRC were compared with HC. The diagnostic performance (area under curve, AUC) of the deep learning algorithm was 0.84 for IBD and 0.80 for CRC prediction. Based on the rat fecal microbiome data, IBD and CRC risks were increased in HFD-fed rats versus RCD-fed rats. Interestingly, in the HFD-induced obesity model, the IBD and CRC risk scores were significantly lowered by the administration of ethanol extracts of NF, but not by the administration of water extracts of NF. In conclusion, changes in the fecal microbiome of obesity by Western diet could be important risk factors for the development of IBD and CRC. The risk prediction model developed in this study could be used to evaluate dietary efficacy.
... So far, the vast knowledge about the effect of gluten/wheat restricted diets on the gut microenvironment is obtained by studies on healthy individuals. These studies have described the ability of a gluten-free diet to shape the gut microbiota composition (Bonder et al., 2016;De Palma et al., 2009;Hansen et al., 2018;Pinto-Sanchez et al., 2020) and modulate activity levels of bacterial metabolic pathways (Bonder et al., 2016). Additionally, a low gluten diet seems to induce moderate changes in the gut microbiota (eg. ...
Article
Full-text available
Irritable bowel syndrome (IBS) is a disorder of gut–brain interaction with a complex pathophysiology. Growing evidence suggests that alterations of the gut microenvironment, including microbiota composition and function, may be involved in symptom generation. Therefore, attempts to modulate the gut microenvironment have provided promising results as an indirect approach for IBS management. Antibiotics, probiotics, prebiotics, food and faecal microbiota transplantation are the main strategies for alleviating IBS symptom severity by modulating gut microbiota composition and function (eg. metabolism), gut barrier integrity and immune activity, although with varying efficacy. In this narrative review, we aim to provide an overview of the current approaches targeting the gut microenvironment in order to indirectly manage IBS symptoms.
... Butyrate produced by intestinal microbes induces expression of hypoxia-inducible factor (HIF) to maintain the integrity of intestinal barriers, thereby mitigating IBD symptoms (105). A population study in middle-aged Danish adults consuming a low-gluten diet revealed that qualitative changes in dietary fiber induced moderate changes in the intestinal microbiome (106). Furthermore, high dietary fiber improved the expression of TJ proteins [e.g., zonula occludens (ZO), occludin, and claudin], which is inhibited by dextran sulfate sodium (DSS), and total SCFAs. ...
Article
Full-text available
Inflammatory bowel disease (IBD) is an idiopathic inflammatory disease. Environmental sanitization, modern lifestyles, advanced medicines, ethnic origins, host genetics and immune systems, mucosal barrier function, and the gut microbiota have been delineated to explain how they cause mucosal inflammation. However, the pathogenesis of IBD and its therapeutic targets remain elusive. Recent studies highlighted the importance of the human gut microbiota in health and disease, suggesting that the pathogenesis of IBD is highly associated with imbalances of the gut microbiota or alterations of epithelial barrier function in the gastrointestinal (GI) tract. Moreover, diet-induced alterations of the gut microbiota in the GI tract modulate immune responses and perturb metabolic homeostasis. This review summarizes recent findings on IBD and its association with diet-induced changes in the gut microbiota, furthermore, discusses how diets can modulate host gut microbes and immune systems, potentiating the impact of personalized diets on therapeutic targets for IBD.
Article
A gluten-free diet is defined as a dietary model that does not include wheat, barley, rye and their products. It is a therapeutic approach that should be applied for life in individuals with a diagnosis of celiac or gluten intolerance. However, recently, healthy individuals can be applied frequently to reduce complaints such as flatulence, bloating and body weight loss. Although the positive or negative effects of this practice are not clear, the effects on the microbiota are being examined. Gut microbiota is defined as an organ that affects human health by producing harmful or beneficial metabolites and protects against pathogens, contributes to the immune system and regulates gastrointestinal development. It has been reported that the gluten-free diet partly or exactly corrects the dysbiosis in the microbiota in individuals who need to follow a gluten-free diet, such as celiac patients. It is also stated that probiotic supplementation in addition to the gluten-free diet in celiac disease helps to restore intestinal homeostasis. On the contrary, it has been reported that gluten-free diet practice causes microbiota dysbiosis in healthy individuals. It is emphasized that this effect is related to the low ingredient of complex carbohydrates and vegetable protein and high fat in gluten-free diet. As a result, gluten-free diet is effective on the intestinal microbiota, and effect on those with celiac disease and healthy individuals may be different. Especially in studies with healthy individuals, there are limitations due to the small sample size and short practice times. More studies are needed to reach a clear mechanism and data in the evaluation of the effect of gluten on the microbiota.
Article
The human gut microbiome responds dynamically to changes in diet, contributing to the diversity of the human diet. The composition of the diet can also influence the metabolic products that are ultimately present in the intestinal lumen and plasma. Observations in humans on vegan, omnivorous, or synthetic enteral diets deficient in fiber have demonstrated dramatic imbalances in the gut microbiome, accompanied by changes in carbohydrate and amino acid-derived bacterial metabolites. It has been established that diet plays a dominant role in the formation of interindividual variations in host-associated microbial communities. Given the impact of the microbial community on human health and its multiple mechanisms of action, establishing the impact of different dietary approaches and models is critical.
Article
Full-text available
Several gastrointestinal phenotypes and impairment of duodenal mucosal barrier have been reported in clinical studies in patients with functional dyspepsia (FD). Due to the preferential colonization of the mucosa, intestinal microbes and their metabolites are commonly involved in host metabolism and immune responses. However, there are no studies on the intertwined correlation among multi-level data. For more comprehensive illustrating, a multi-omics analysis focusing on the duodenum was performed in the FD rat model. We found that differential microbiomes in the duodenum were significantly correlated with the biosynthesis of lipopolysaccharide and peptidoglycan. The innate immune response-related genes, which were upregulated in the duodenum, were associated with the TLR2/TLR4-NFκB signaling pathway. More importantly, arachidonyl ethanolamide (anandamide, AEA) and endocannabinoid analogues showed linear relationships with the FD phenotypes. Taken together, multi-level data from microbiome, transcriptome and metabolome reveal that AEA may regulate duodenal low-grade inflammation in FD. These results suggest an important cue of gut microbiome–endocannabinoid system axis in the pathogenesis of FD.
Article
Full-text available
In this review after a lifelong research career, my personal opinion on the development of type 1 diabetes (T1D) from its very start to clinical manifestation will be described. T1D is a disease of an increased intestinal permeability and a reduced pancreas volume. I am convinced that virus might be the initiator and that this virus could persist on strategically significant locations. Furthermore, intake of gluten is important both in foetal life and at later ages. Disturbances in sphingolipid metabolism may also be of crucial importance. During certain stages of T1D, T cells take over resulting in the ultimate destruction of beta cells, which manifests T1D as an autoimmune disease. Several preventive and early treatment strategies are mentioned. All together this review has more new theories than usually, and it might also be more speculative than ordinarily. But without new ideas and theories advancement is difficult, even though everything might not hold true during the continuous discovery of the etiology and pathogenesis of T1D.
Article
Full-text available
Wheat is an important staple grain for humankind globally because of its end-use quality and nutritional properties and its adaptability to diverse climates. For a small proportion of the population, specific wheat proteins can trigger adverse immune responses and clinical manifestations such as celiac disease, wheat allergy, baker's asthma, and wheat-dependent exercise-induced anaphylaxis (WDEIA). Establishing the content and distribution of the immuno-stimulatory regions in wheat has been hampered by the complexity of the wheat genome and the lack of complete genome sequence information. We provide novel insights into the wheat grain proteins based on a comprehensive analysis and annotation of the wheat prolamin Pfam clan grain proteins and other non-prolamin allergens implicated in these disorders using the new International Wheat Genome Sequencing Consortium bread wheat reference genome sequence, RefSeq v1.0. Celiac disease and WDEIA genes are primarily expressed in the starchy endosperm and show wide variation in protein-and transcript-level expression in response to temperature stress. Nonspecific lipid transfer proteins and -amylase trypsin inhibitor gene families, implicated in baker's asthma, are primarily expressed in the aleurone layer and transfer cells of grains and are more sensitive to cold temperature. The study establishes a new reference map for immunostimulatory wheat proteins and provides a fresh basis for selecting wheat lines and developing diagnostics for products with more favorable consumer attributes.
Article
Full-text available
Introduction Increased gut permeability (“leaky gut”) has been proposed as a potential contributor to age-related inflammation and gut dysbiosis. However, information on the relationship between a leaky gut and inflammation and physical frailty during aging are limited. Objective To investigate the hypothesis that an aging-associated leaky gut is linked to the age-related inflammation and frailty. Methods Two cohorts of healthy adults were studied: young (18–30 years old, n = 19) and older (≥70 years old, n = 18). Serum concentrations of the tumor necrosis factor (TNF)-α and interleukin (IL)-6, zonulin (a marker for leaky gut), and high-mobility group box protein (HMGB1, a nuclear protein triggering inflammation) were measured. Correlations of serum levels of zonulin and HMGB1 with strength of plantar flexor muscles and number of steps taken per day were analyzed. Results Serum concentration of zonulin and HMGB1 were 22% (P = .005) and 16% (P = .010) higher in the older versus young adults. Serum zonulin was positively associated with concentrations of TNF-α (r = 0.357, P = .032) and IL-6 (r = 0.345, P = .043). Importantly, both zonulin and HMGB1 were negatively correlated with skeletal muscle strength (zonulin: r = −0.332, P = .048; HMGB1: r = −0.383, P = .023), and habitual physical activity (zonulin: r = −0.410, P = .016; HMGB1: r = −0.483, P = .004). Conclusions Serum zonulin was associated with both systemic inflammation and 2 key indices of physical frailty. These data suggest that a leaky gut may play a critical role in the development of age-related inflammation and frailty.
Article
Full-text available
Short-chain fatty acids (SCFA), formed by microbial fermentation, are believed to be involved in the aetiology of obesity and diabetes. This study investigated the effects of colonic administration of physiologically relevant SCFA mixtures on human substrate and energy metabolism. In this randomized, double-blind, crossover study, twelve normoglycaemic men (BMI 25–35 kg/m²) underwent four investigational days, during which SCFA mixtures (200 mmol/L) high in either acetate (HA), propionate (HP), butyrate (HB) or placebo (PLA) were rectally administered during fasting and postprandial conditions (oral glucose load). Before and for two hours after colonic infusions, indirect calorimetry was performed and blood samples were collected. All three SCFA mixtures increased fasting fat oxidation (P < 0.01), whilst resting energy expenditure increased after HA and HP compared with PLA (P < 0.05). In addition, all three SCFA mixtures increased fasting and postprandial plasma peptide YY (PYY) concentrations, and attenuated fasting free glycerol concentrations versus PLA (P < 0.05). Colonic infusions of SCFA mixtures, in concentrations and ratios reached after fibre intake, increased fat oxidation, energy expenditure and PYY, and decreased lipolysis in overweight/obese men. Human intervention studies are warranted to investigate whether these effects translate into long-term benefits for body weight control and insulin sensitivity in the obese insulin resistant state.
Article
Full-text available
Objective To examine the association of long term intake of gluten with the development of incident coronary heart disease. Design Prospective cohort study. Setting and participants 64 714 women in the Nurses’ Health Study and 45 303 men in the Health Professionals Follow-up Study without a history of coronary heart disease who completed a 131 item semiquantitative food frequency questionnaire in 1986 that was updated every four years through 2010. Exposure Consumption of gluten, estimated from food frequency questionnaires. Main outcome measure Development of coronary heart disease (fatal or non-fatal myocardial infarction). Results During 26 years of follow-up encompassing 2 273 931 person years, 2431 women and 4098 men developed coronary heart disease. Compared with participants in the lowest fifth of gluten intake, who had a coronary heart disease incidence rate of 352 per 100 000 person years, those in the highest fifth had a rate of 277 events per 100 000 person years, leading to an unadjusted rate difference of 75 (95% confidence interval 51 to 98) fewer cases of coronary heart disease per 100 000 person years. After adjustment for known risk factors, participants in the highest fifth of estimated gluten intake had a multivariable hazard ratio for coronary heart disease of 0.95 (95% confidence interval 0.88 to 1.02; P for trend=0.29). After additional adjustment for intake of whole grains (leaving the remaining variance of gluten corresponding to refined grains), the multivariate hazard ratio was 1.00 (0.92 to 1.09; P for trend=0.77). In contrast, after additional adjustment for intake of refined grains (leaving the variance of gluten intake correlating with whole grain intake), estimated gluten consumption was associated with a lower risk of coronary heart disease (multivariate hazard ratio 0.85, 0.77 to 0.93; P for trend=0.002). Conclusion Long term dietary intake of gluten was not associated with risk of coronary heart disease. However, the avoidance of gluten may result in reduced consumption of beneficial whole grains, which may affect cardiovascular risk. The promotion of gluten-free diets among people without celiac disease should not be encouraged.
Article
The role of tryptophan-kynurenine metabolism in psychiatric disease is well established, but remains less explored in peripheral tissues. Exercise training activates kynurenine biotransformation in skeletal muscle, which protects from neuroinflammation and leads to peripheral kynurenic acid accumulation. Here we show that kynurenic acid increases energy utilization by activating G protein-coupled receptor Gpr35, which stimulates lipid metabolism, thermogenic, and anti-inflammatory gene expression in adipose tissue. This suppresses weight gain in animals fed a high-fat diet and improves glucose tolerance. Kynurenic acid and Gpr35 enhance Pgc-1α1 expression and cellular respiration, and increase the levels of Rgs14 in adipocytes, which leads to enhanced beta-adrenergic receptor signaling. Conversely, genetic deletion of Gpr35 causes progressive weight gain and glucose intolerance, and sensitizes to the effects of high-fat diets. Finally, exercise-induced adipose tissue browning is compromised in Gpr35 knockout animals. This work uncovers kynurenine metabolism as a pathway with therapeutic potential to control energy homeostasis.
Article
Although humans have cospeciated with their gut-resident microbes, it is difficult to infer features of our ancestral microbiome. Here, we examine the microbiome profile of 350 stool samples collected longitudinally for more than a year from the Hadza hunter-gatherers of Tanzania. The data reveal annual cyclic reconfiguration of the microbiome, in which some taxa become undetectable only to reappear in a subsequent season. Comparison of the Hadza data set with data collected from 18 populations in 16 countries with varying lifestyles reveals that gut community membership corresponds to modernization: Notably, the taxa within the Hadza that are the most seasonally volatile similarly differentiate industrialized and traditional populations. These data indicate that some dynamic lineages of microbes have decreased in prevalence and abundance in modernized populations.
Article
Background & Aims Dietary restriction of fermentable carbohydrates (a low FODMAP diet) has been reported to reduce symptoms in some patients with irritable bowel syndrome (IBS). We performed a randomized, placebo-controlled study to determine its effects on symptoms and the fecal microbiota in patients with IBS. Methods We performed a 2x2 factorial trial of 104 patients with IBS (18–65 years old), based on the Rome III criteria, at 2 hospitals in the United Kingdom. Patients were randomly assigned (blinded) to groups given counselling to follow a sham diet or diet low in FODMAPs for 4 weeks, along with a placebo or probiotic supplement (VSL#3), resulting in 4 groups (27 receiving sham diet/placebo, 26 receiving sham diet/probiotic, 24 receiving low FODMAP diet /placebo, and 27 receiving low FODMAP diet /probiotic). The sham diet restricted a similar number of staple and non-staple foods as the low FODMAP diet; the diets had similar degrees of difficulty to follow. Dietary counselling was given to patients in all groups and data on foods eaten and compliance were collected. The incidence and severity of 15 gastrointestinal symptoms and overall symptoms were measured daily for 7 days before the study period; along with stool frequency and consistency. At baseline, global and individual symptoms were measured, along with generic and disease-specific health-related quality of life, using standard scoring systems. All data were collected again at 4 weeks, and patients answered questions about adequate symptom relief. Fecal samples were collected at baseline and after 4 weeks and analyzed by quantitative PCR and 16S rRNA sequencing. The co-primary endpoints were adequate relief of symptoms and stool Bifidobacterium species abundance at 4 weeks. Results There was no significant interaction between the interventions in adequate relief of symptoms (P=.52) or Bifidobacterium species (P=.68). In the intention-to-treat analysis, a higher proportion of patients in the low FODMAP diet had adequate symptom relief (57%) vs than in the sham diet group (38%), although the difference was not statistically significant (P=.051). In the per-protocol analysis, a significantly higher proportion of patients on the low FODMAP diet had adequate symptom relief (61%) than in the sham diet group (39%) (P=.043). Total mean IBS- Severity Scoring System score was significantly lower for patients on the low FODMAP diet (173±95) than the sham diet (224 ± 89)(P=.001), but not different between those given probiotic (207 ± 98) or placebo (192 ± 93)(P=.721) Abundance of Bifidobacterium species was lower in fecal samples from patients on the low FODMAP diet (8.8 rRNA genes/g) than patients on the sham diet (9.2 rRNA genes/g) (P=.008), but higher in patients given probiotic (9.1 rRNA genes/g) than patients given placebo (8.8 rRNA genes/g) (P=.019). There was no effect of the low FODMAP diet on microbiota diversity in fecal samples. Conclusions In a placebo-controlled study of patients with IBS, a low FODMAP diet associates with adequate symptom relief and significantly reduced symptom scores compared with placebo. It is not clear whether changes resulted from collective FODMAP restriction or removal of a single component, such as lactose. Co-administration of the probiotic VSL#3 increased numbers of Bifidobacterium species, compared with placebo, and might be given to restore these bacteria to patients on a low FODMAP diet. Trial registration no: ISRCTN02275221.
Article
Objective: The effects of dietary interventions on gut bacteria are ambiguous. Following a previous intervention study, we aimed to determine how differing diets impact gut bacteria and if bacterial profiles predict intervention response. Design: Sixty-seven patients with IBS were randomised to traditional IBS (n=34) or low fermentable oligosaccharides, disaccharides, monosaccharides and polyols (FODMAPs) (n=33) diets for 4 weeks. Food intake was recorded for 4 days during screening and intervention. Faecal samples and IBS Symptom Severity Score (IBS-SSS) reports were collected before (baseline) and after intervention. A faecal microbiota dysbiosis test (GA-map Dysbiosis Test) evaluated bacterial composition. Per protocol analysis was performed on 61 patients from whom microbiome data were available. Results: Responders (reduced IBS-SSS by ≥50) to low FODMAP, but not traditional, dietary intervention were discriminated from non-responders before and after intervention based on faecal bacterial profiles. Bacterial abundance tended to be higher in non-responders to a low FODMAP diet compared with responders before and after intervention. A low FODMAP intervention was associated with an increase in Dysbiosis Index (DI) scores in 42% of patients; while decreased DI scores were recorded in 33% of patients following a traditional IBS diet. Non-responders to a low FODMAP diet, but not a traditional IBS diet had higher DI scores than responders at baseline. Finally, while a traditional IBS diet was not associated with significant reduction of investigated bacteria, a low FODMAP diet was associated with reduced Bifidobacterium and Actinobacteria in patients, correlating with lactose consumption. Conclusions: A low FODMAP, but not a traditional IBS diet may have significant impact on faecal bacteria. Responsiveness to a low FODMAP diet intervention may be predicted by faecal bacterial profiles. Trial registration number: NCT02107625.