Determinants of Reduced Genetic Capacity for Butyrate Synthesis by the Gut Microbiome in Crohn’s Disease and Ulcerative Colitis

Determinants of Reduced Genetic Capacity for Butyrate Synthesis by the Gut Microbiome in... Abstract Background and Aims Alterations in short chain fatty acid metabolism, particularly butyrate, have been reported in inflammatory bowel disease, but results have been conflicting because of small study numbers and failure to distinguish disease type, activity or other variables such as diet. We performed a comparative assessment of the capacity of the microbiota for butyrate synthesis, by quantifying butyryl-CoA:acetate CoA-transferase [BCoAT] gene content in stool from patients with Crohn’s disease [CD; n = 71], ulcerative colitis [UC; n = 58] and controls [n = 75], and determined whether it was related to active vs inactive inflammation, microbial diversity, and composition and/or dietary habits. Methods BCoAT gene content was quantified by quantitative polymerase chain reaction [qPCR]. Disease activity was assessed clinically and faecal calprotectin concentration measured. Microbial composition was determined by sequencing 16S rRNA gene. Dietary data were collected using an established food frequency questionnaire. Results Reduced butyrate-synthetic capacity was found in patients with active and inactive CD [p < 0.001 and p < 0.01, respectively], but only in active UC [p < 0.05]. In CD, low BCoAT gene content was associated with ileal location, stenotic behaviour, increased inflammation, lower microbial diversity, greater microbiota compositional change, and decreased butyrogenic taxa. Reduced BCoAT gene content in patients with CD was linked with a different regimen characterised by lower dietary fibre. Conclusions Reduced butyrate-synthetic capacity of the microbiota is more evident in CD than UC and may relate to reduced fibre intake. The results suggest that simple replacement of butyrate per se may be therapeutically inadequate, whereas manipulation of microbial synthesis, perhaps by dietary means, may be more appropriate. Butyrate synthesis, inflammatory bowel disease, microbiota 1. Introduction The inflammatory bowel diseases [IBD], Crohn’s disease [CD] and ulcerative colitis [UC], are heterogeneous and multifactorial disorders involving variable influences including genetic predisposition, immune-mediated tissue damage, and environmental modifiers, of which the gut microbiota may be the proximate environmental influence.1 Prominent among the microbial metabolites that interact with the host are short chain fatty acids [SCFA]. These are normally produced by the colonic microbiota by fermentation of complex undigested polysaccharides such as dietary fibres. The SCFA that has been most plausibly linked with IBD is butyrate, which can be synthesised from butyryl-CoA by two different enzymes: butyrate kinase and butyryl-CoA:acetate CoA-transferase [BCoAT], the latter being predominant in the human colonic ecosystem.2 Butyrate provides energy to the colonocytes, promoting epithelial cell growth. It also acts as a signalling molecule regulating cell proliferation and differentiation. Butyrate has been described to possess anti-carcinogenic, anti-lipogenic and anti-inflammatory potential, and may play a role in satiety, insulin sensitivity, oxidative stress, motility, sodium and water absorption, stimulation of mucous secretion, and increase of vascular flow.3–5 Butyrate inhibits inflammatory responses by decreasing proinflammatory cytokine expression via inhibition of NF-κB transcription factor activation in immune cells,6 and is considered a potent communicator to the immune system, acting mainly as an inhibitor of histone deacetylases. In that way, it favours efficient generation of regulatory T cells7,8 and homeostasis.9 Therefore, butyrate has been employed in several randomised clinical trials and interventional studies to analyse its efficacy for symptom relief, especially in diseases involving inflammation.10 Disturbed butyrate metabolism has been linked with IBD, but results have been either discrepant or conflicting in relation to CD vs UC, predominantly based on faecal levels, and seldom related to disease activity.11–14 Diminished colonic butyrogenic bacteria in IBD have been reported in faecal and mucosal samples from patients with IBD15 and most notably include Faecalibacterium prausnitzii16,17 and the Roseburia genus,18 which were also decreased in new-onset CD.19 In addition, lower numbers of the genus Butyricicoccus were reported in stools of patients with IBD.20 Alterations in butyrate production in IBD have been also suggested by metagenomic analysis.21,22 We hypothesised that disturbances in butyrogenic capacity of colonic microbiota may be more informative than faecal butyrate levels. However, direct quantification of the capacity of the gut microbiota to produce butyrate across active and inactive stages has not been performed. We have addressed this by quantifying BCoAT gene content in faecal microbiota from patients with CD and UC in comparison with controls, and results were related to clinical variables including disease activity and habitual diet. 2. Material and Methods 2.1. Study population Patients with CD and UC were recruited from dedicated IBD specialty clinics at Cork University Hospital and Mercy University Hospital, Cork, Ireland. Eligible patients were aged 18–67 years with a well-established diagnosis of chronic recurrent CD or UC. The following exclusion criteria were applied in patient recruitment: pregnant female; a history of total colectomy; significant acute or chronic coexisting illness; treatment with experimental drugs; and malignant or any concomitant end-stage organ disease. Healthy volunteers were recruited from the same geographical region and mostly among University College Cork staff. Informed consent was obtained from all study subjects and the study was approved by the Clinical Research Ethics Committee of the Cork Teaching Hospitals, University College Cork, National University of Ireland. All IBD subjects completed a clinical interview with a gastroenterologist and surveys detailing their demographic [age, gender, nationality, body mass index, smoking, breastfeeding, and born by caesarean section] and clinical characteristics [history of IBD-related surgeries, frequency of symptoms, IBD clinical indices, current medication, perceived stress, and history of other diseases and surgeries aside from IBD]. Patients with active IBD were encouraged to be enrolled so as to achieve a wide representation of disease activities in the studied cohort, with some of them having a change to a more aggressive treatment when recruited. Demographic characteristics were also obtained for control individuals. In addition, all subjects [except for eight patients with CD, two patients with UC, and eight control individuals] completed a detailed food frequency questionnaire [FFQ], as previously described by Claesson et al.23 Subjects were instructed to store faecal samples in a refrigerator from which they were transported within 24 h to the research laboratory and stored at 4°C. Samples were subdivided into aliquots of approximately 0.2 g and 0.5 g within 24 h of arrival for further DNA extraction and calprotectin measurement, respectively, and then frozen at -80°C. Wet weight of faecal samples was measured in those aliquots intended to be for DNA extraction before being frozen. Diarrhoeal samples with liquid consistency were annotated as such. 2.2. 16S rRNA gene sequencing and quantitative polymerase chain reaction assay A total number of 204 stool samples were included in this study [71 CD and 58 UC patients, and 75 healthy controls]. Some of the samples from diseased individuals were classified into two sub-groups [inactive and active] when fulfilling specific criteria: CD inactive when Harvey-Bradshaw Index [HBI]24 ≤ 4 and symptom activity described as ‘occasional [a day or more once a month]’, or ‘rare’, or ‘absent’; CD active when HBI ≥ 5 and symptom activity categorised as ‘active all the time’ or ‘active often [some days in each week]’; UC inactive when Powell-Tuck Index [PTI]25 ≤ 3 and Ulcerative Colitis Activity Index [UCAI]26 ≤ 4 and symptom activity described as ‘occasional’, or ‘rare’ or ‘absent’; and UC active when PTI ≥ 4 and UCAI ≥ 5 and symptom activity categorised as ‘active all the time’ or ‘active often’. According to these criteria, the number of patients in each sub-group was 16 CD active, 34 CD inactive, 19 UC active and 27 UC inactive. DNA extraction on faecal samples was carried out as previously described in Clooney et al.,27 using the QIAamp Fast DNA stool kit [Qiagen GmbH, Hilden, Germany] including repeated bead-beating. Library preparation of 16S rRNA gene amplicon for sequencing was performed following the Illumina [San Diego, CA, USA] recommendations with some modifications described in the aforementioned study.27 Primers for amplification of the V3-V4 hypervariable region of the 16S rRNA gene were selected from Klindworth et al.28 MiSeq sequencing of the 16S library was made at Eurofins Genomics [Ebersberg, Germany]. The number of copies of BCoAT gene was determined in genomic DNA [gDNA] samples [previously diluted 1:5] by quantitative polymerase chain reaction [qPCR]. Degenerate primers designed from the BCoAT sequence in four butyrate-producer species,29 BCoATscrF [GCIGAICATTTCACITGGAAYWSITGGCAYATG] and BCoATscrR [CCTGCCTTTGCAATRTCIACRAANGC] [Eurofins Genomics], were employed for amplification. The qPCR reaction was performed in a total volume of 20 µL containing LightCycler SYBR Green I Master [Roche, Penzberg, Germany], nuclease-free water, primers at a final concentration of 2 µM, and 2 µL of gDNA. Amplification was carried out in a LightCycler 480 Instrument II [Roche] under the following conditions: 95°C for 10 min, 40 cycles of 95°C for 10 s, 53°C for 15 s, and 72°C for 30 s with data acquisition at 72°C, and a stepwise increase of the temperature to 95°C [at 0.29°C/s] with continuous acquisition to obtain melting curve data. All samples were run in duplicates and mean values were used in calculations. To generate standard curves, calculated amounts of the linearised plasmid pGEM T Easy [Promega, Madison, WI, USA], in which the amplified region from the BCoAT gene had been inserted, were used. Plasmid concentration was measured using Nano-Drop 2000 Spectrophotometer [Thermo Scientific, Wilmington, DE, USA], and the number of copies was calculated using the DNA Copy Number and Dilution Calculator tool at Life Technologies website. Serial dilutions [from 107 to 10] were amplified in duplicates to create the standard curve for extrapolation of BCoAT gene copies using the Ct values of the samples. Results were multiplied according to the dilution factor, normalised as copies per gram of wet weight of faeces and, finally, expressed as log10 values. No diarrhoeal samples presenting liquid consistency were included in the study to avoid false reductions of bacterial load due to increased percentage of faecal water. Repeatability conditions were tested in a sample extracted six different times and processed in different qPCR assays. Coefficient of variation for these six samples was 1.5%, showing a good reproducibility between different DNA extractions and different qPCR runs. 2.3. Bioinformatics analysis A mean of 19464 (± 4869 standard deviation [SD]) reads per sample was obtained. Adaptor sequences were removed from V3-V4 16S rRNA gene amplicon reads using cutadapt with a 0.2 error rate allowance. The script fastq_merge [USEARCH version 7.0.1090] was employed to merge forward and reverse paired reads. Demultiplexing was carried out using split_libraries [QIIME] and only reads in the range of 390 bp to 465 bp in length and with an average PHRED score of Q25 and above were retained for downstream analysis. Operational taxonomic units [OTUs] were clustered using a de novo approach [QIIME] and removal of chimeric sequences was carried out using uchime_ref in conjunction with the ChimeraSlayer GOLD database.30 The OTU table was normalised using the variance stabilising technique [VST] from the DESeq R library. Taxonomy [phylum to genus] was assigned to OTUs using the Mothur implementation of the Ribosomal Database Project [RDP] classifier [version 11.4].31 Any sequence with less than an 80% bootstrap value was assigned as unclassified at that particular rank. SPINGO, with default parameters, was used for species and Clostridium cluster assignment on demultiplexed and quality filtered reads.32 2.4. Faecal calprotectin assay A -80°C frozen aliquot for each sample, containing approximately 0.5 g of faecal material, was used to measure calprotectin concentration with EliATM Calprotectin Immunoassay v2 [Thermo Scientific, Uppsala, Sweden], following manufacturer’s instructions. First, calprotectin was extracted from faeces using EliA Stool Extraction Kit. Samples were picked employing a rod with four notches that indicated the stool quantity required, and were introduced in the extraction buffer. Subsequently, they were vortexed and centrifuged. Finally, the resulting supernatant was processed in an ImmunoCap 250 autoanalyser [Thermo Scientific] set up with the buffers, conjugates, and reagents for EliATM calprotectin assay. A control curve was measured for each calprotectin run as a quality control. Results were expressed as µg/g faeces. Results below detection limit [< 3.8 µg/g] were accounted as 4 µg/g in calculations. Results above lineal range limit [> 6000 µg/g] were considered as 6000 µg/g in calculations. 2.5. Statistical analysis Normality distribution was assessed using the Kolmogorov-Smirnov test [when n ≥ 50] or the Shapiro-Wilk test [when n < 50]. Logarithmic transformation was applied to calprotectin when normality was not assumed. In cases where normality could not be assumed, non-parametric tests were employed [Kruskal-Wallis for multiple comparison and Mann-Whitney for independent comparisons]. Alpha and beta diversity were calculated using the phyloseq library, followed by Principal Coordinate Analysis [PCoA] graphed via the ggplot2 library. Wilcoxon rank sum tests were used to compare alpha and beta diversity between groups and taxa [after normalisation via proportions]. Principal component analysis [PCA] was performed to compare the categorical diet data. Procrustes [ade4 R library] was employed to access correlations between the beta diversity [weighted UniFrac with VST normalisation] and dietary data [PCA]. All p-values were adjusted where necessary using the Benjamini and Hochberg method to correct multiple testing. For linear regression analysis, Spearman correlations were used. Statistical analyses were performed in GraphPad Prism v5, SPSS v15, and R v3.3.0. 3. Results 3.1. Butyryl-CoA:acetate CoA-transferase gene content of microbiota in Crohn’s and ulcerative colitis No major differences for demographic data were found between patients with CD and UC except for previous IBD surgery and 5-aminosalicylic acid compounds usage [Table 1]. Control individuals presented a lower percentage of obese subjects and higher percentage of individuals who had been breastfed and born by caesarean section than patients with IBD. They were also statistically significantly younger than UC patients. Table 1. Clinical profile of study populations. Montreal classification for patients with Crohn’s disease [CD] was collected for 58 subjects. Numbers in parenthesis for medication and Montreal classification correspond to absolute numbers.   CD  UC  Controls  n  71  58  75  Age [years]a  44.3 ± 12.3  48.9 ± 12.0  41.2 ± 16.2  % male  54.9%  65.5%  46.6%  Years since diagnosis  14.2 ± 9.7  16.6 ± 10.1  -  % Previous IBD surgeryb  49.3  0.0  -  Body mass indexc,d  27.7 ± 5.5  26.9 ± 4.3  24.7 ± 3.7  % Obese [BMI > 30]e  27.5  21.4  6.9  % Smokers  8.5  8.5  12.9  % Breastfedf  31.1  32.5  54.0  % Caesarean sectionf  6.5  0.0  14.5  K6 stress scale [K6 ≤ 18]  25.4 ± 4.4 [9%]  26.6 ± 3.1 [4%]  -  Inactive phase  34  27  -  Active phase  16  19  -  Medication  % 5-ASA compoundsb  39.4 [28]  72.4 [42]  -  % Steroids  16.9 [12]  13.8 [8]  -  % Immunosuppressants  25.4 [18]  27.6 [16]  -  % Biologics  28.2 [20]  17.2 [10]  -  Montreal classification  % Age at diagnosis  A1: 6.9 [4]  -  -  A2: 79.3 [46]  A3: 13.8 [8]  % Location  L1: 29.3 [17]  -  -  L2: 25.9 [15]  L3: 43.1 [25]  L1/L4: 1.7 [1]  % Behaviour  B1: 43.1 [25]  -  -  B2: 32.8 [19]  B3: 24.1 [14]  P: 16.7 [9]    CD  UC  Controls  n  71  58  75  Age [years]a  44.3 ± 12.3  48.9 ± 12.0  41.2 ± 16.2  % male  54.9%  65.5%  46.6%  Years since diagnosis  14.2 ± 9.7  16.6 ± 10.1  -  % Previous IBD surgeryb  49.3  0.0  -  Body mass indexc,d  27.7 ± 5.5  26.9 ± 4.3  24.7 ± 3.7  % Obese [BMI > 30]e  27.5  21.4  6.9  % Smokers  8.5  8.5  12.9  % Breastfedf  31.1  32.5  54.0  % Caesarean sectionf  6.5  0.0  14.5  K6 stress scale [K6 ≤ 18]  25.4 ± 4.4 [9%]  26.6 ± 3.1 [4%]  -  Inactive phase  34  27  -  Active phase  16  19  -  Medication  % 5-ASA compoundsb  39.4 [28]  72.4 [42]  -  % Steroids  16.9 [12]  13.8 [8]  -  % Immunosuppressants  25.4 [18]  27.6 [16]  -  % Biologics  28.2 [20]  17.2 [10]  -  Montreal classification  % Age at diagnosis  A1: 6.9 [4]  -  -  A2: 79.3 [46]  A3: 13.8 [8]  % Location  L1: 29.3 [17]  -  -  L2: 25.9 [15]  L3: 43.1 [25]  L1/L4: 1.7 [1]  % Behaviour  B1: 43.1 [25]  -  -  B2: 32.8 [19]  B3: 24.1 [14]  P: 16.7 [9]  IBD, inflammatory bowel disease; UC, ulcerative colitis; BMI, body mass index; 5-ASA, 5-aminosalicylic acid. ap < 0.01 for UC vs controls. bp < 0.001 for CD vs UC. cp < 0.01 for CD vs controls. dp < 0.05 for UC vs controls. ep < 0.01. fp < 0.05. View Large The number of copies of BCoAT gene determined by qPCR did not correlate with age for the whole population or within each group. We observed no correlation with time elapsed after CD/UC diagnosis, either. Although no differences in BCoAT gene content were found between males and females when analysing all samples, there were significant higher BCoAT gene levels for males in UC. Regarding disease, BCoAT gene content was significantly lower in CD samples compared with healthy controls and UC [p < 0.001 and p < 0.05, respectively], whereas no differences were detected between control individuals and UC [Figure 1A; and Supplementary Table 1, available as Supplementary data at ECCO-JCC online]. When the disease activity was considered, BCoAT gene content remained significantly different between controls and both CD groups and also between controls and the UC active group, whereas patients with inactive UC showed higher BCoAT gene levels than both CD groups [Figure 1B; and Supplementary Table 1]. Similar differences between groups were observed when patients with IBD were classified according to faecal calprotectin levels, a specific biomarker of active intestinal inflammation, though no change in BCoAT gene content was observed between healthy controls and patients with UC and high calprotectin concentration [Figure 1C; and Supplementary Table 1]. As expected, concentrations of faecal calprotectin were significantly higher [p < 0.001] for active patients in both CD [mean ± SD: 971 ± 1652 vs 49 ± 58 µg/g] and UC [mean ± SD: 967 ± 1459 vs 57 ± 74 µg/g], demonstrating a good concordance with patient classification following clinical criteria [Figure 1D]. Figure 1. View largeDownload slide Quantification of BCoAT gene content in faecal samples of patients with IBD and healthy controls by qPCR. Tukey boxplot [whiskers represent median ± 1.5 interquartile range] when three groups were considered: CD [n = 71], controls [n = 75], and UC [n = 58] [A]; when patients with IBD were sub-grouped according to disease activity: CD active [n = 16], CD remission [n = 34], controls [n = 75], UC remission [n = 27] and UC active [n = 19] [B]; and when patients with IBD were sub-grouped according to faecal calprotectin [FCAL] concentration: CD and FCAL > 100 [n = 29], CD and FCAL < 100 [n = 42], controls [n = 75], UC and FCAL < 100 [n = 36], and UC and FCAL > 100 [n = 22]. Graph depicting mean calprotectin concentration with whiskers representing standard error of the mean for patients with CD and UC classified according to disease activity [D]. Tukey boxplot for BCoAT gene content in patients with CD classified according to disease location [L1 = ileal, L2 = colonic, L3 = ileo-colonic] [E]; and disease behaviour [B1 = non-stenotic/non-penetrating, B2 = stenotic, B3 = penetrating] [F]. *p < 0.05, **p < 0.01, ***p < 0.001. IBD, inflammatory bowel disease; qPCR, quantitative polymerase chain reaction; CD, Crohn’s disease; UC, ulcerative colitis. Figure 1. View largeDownload slide Quantification of BCoAT gene content in faecal samples of patients with IBD and healthy controls by qPCR. Tukey boxplot [whiskers represent median ± 1.5 interquartile range] when three groups were considered: CD [n = 71], controls [n = 75], and UC [n = 58] [A]; when patients with IBD were sub-grouped according to disease activity: CD active [n = 16], CD remission [n = 34], controls [n = 75], UC remission [n = 27] and UC active [n = 19] [B]; and when patients with IBD were sub-grouped according to faecal calprotectin [FCAL] concentration: CD and FCAL > 100 [n = 29], CD and FCAL < 100 [n = 42], controls [n = 75], UC and FCAL < 100 [n = 36], and UC and FCAL > 100 [n = 22]. Graph depicting mean calprotectin concentration with whiskers representing standard error of the mean for patients with CD and UC classified according to disease activity [D]. Tukey boxplot for BCoAT gene content in patients with CD classified according to disease location [L1 = ileal, L2 = colonic, L3 = ileo-colonic] [E]; and disease behaviour [B1 = non-stenotic/non-penetrating, B2 = stenotic, B3 = penetrating] [F]. *p < 0.05, **p < 0.01, ***p < 0.001. IBD, inflammatory bowel disease; qPCR, quantitative polymerase chain reaction; CD, Crohn’s disease; UC, ulcerative colitis. We also performed statistical analysis for BCoAT gene concentrations among the different categories of Montreal classification in patients with CD. No differences were found for age at diagnosis, but we detected significant changes depending on disease location and behaviour. Patients with ileal location showed significantly lower BCoAT gene levels than those with colonic location [Figure 1E; and Supplementary Table 1]. Regarding behaviour, patients with stenotic phenotype had the more decreased BCoAT gene content, which was significantly lower than BCoAT gene values in patients with non-stenotic/non-penetrating disease [Figure 1F; and Supplementary Table 1]. Otherwise, BCoAT gene concentration was not different between patients with CD who underwent surgery and those with no previous surgery. 3.2. Relationship between patients sub-grouped according to BCoAT gene content and treatment, clinical variables, and inflammation To examine further associations with current treatment and with other variables [disease activity, obesity, previous IBD surgery, and being breastfed], patients with CD and UC classified as active or inactive were sub-grouped each into two categories, using a cut-off of 9.5 [log10] number of copies of BCoAT gene/gram wet faeces. This cut-off value was chosen to achieve a similar number of samples in each category: CD ≤ 9.5 [n = 29], CD > 9.5 [n = 21], UC ≤ 9.5 [n = 19], and UC > 9.5 [n = 27]. In that way, BCoAT level was transformed into a dichotomous nominal variable for statistical analysis [Table 2; and Supplementary Figure 1, available as Supplementary data at ECCO-JCC online]. Univariate analysis showed that low BCoAT levels were associated with active CD, whereas no association between BCoAT gene concentration and any other variable was found in patients with UC. We estimated that patients with active CD had 4.9 times higher probability of presenting BCoAT levels below 9.5 log10 copies/g. In a multivariate logistic regression model considering all four categories of treatment, or all four other variables, disease activity remained as the only variable with statistical significance in patients with CD. Treatment with biologics was associated with low BCoAT gene content in the multivariate analysis for patients with UC, but this observation was affected by a low number of anti-tumour necrosis factor [TNF]-treated UC patients in the analysis and the fact that 62.5% of them were in active stage because of being recruited in an early induction phase. Table 2. Univariate and multivariate logistic regression analysis regarding treatment [A] and other data [disease activity, obesity, previous IBD surgery and being breastfed] [B]. Odds ratio [OR], confidence interval [CI] at 95% and p-value [p] in univariate analysis are referred to present BCoAT levels below 9.5 log10 number of copies of BCoAT gene/g wet faeces; p-value for univariate analysis was calculated using Fisher’s exact test. Surgery was not considered in the analysis of patients with UC. Smoking and caesarean section were not considered as presenting low numbers.   Crohn’s disease  Ulcerative colitis  Univariate  Multivariate  Univariate  Multivariate    OR  CI 95%  p  p  OR  CI 95%  p  p  A. Treatment  5-ASA compounds  0.82  0.3–2.6  0.78  0.73  1.18  0.3–4.4  1.00  0.81  Steroids  5.22  0.6–47.1  0.22  0.11  2.13  0.4–10.9  0.42  0.36  Immunosuppressant  2.24  0.6–8.5  0.34  0.23  1.62  0.4–6.1  0.51  0.48  Biologics  1.44  0.4–5.2  0.75  0.57  5.77  1.0–32.7  0.05  0.03  B. Other data                  Disease active  4.88  1.2–20.3  0.03  0.02  1.53  0.5–5.0  0.55  0.07  BMI > 30  0.28  0.1–1.0  0.06  0.14  0.88  0.2–4.3  1.00  0.73  Previous surgery  1.24  0.4–3.9  0.78  0.86  -  -  -  -  Breastfed  1.13  0.3–4.3  1.00  0.41  0.70  0.2–3.2  0.72  0.64    Crohn’s disease  Ulcerative colitis  Univariate  Multivariate  Univariate  Multivariate    OR  CI 95%  p  p  OR  CI 95%  p  p  A. Treatment  5-ASA compounds  0.82  0.3–2.6  0.78  0.73  1.18  0.3–4.4  1.00  0.81  Steroids  5.22  0.6–47.1  0.22  0.11  2.13  0.4–10.9  0.42  0.36  Immunosuppressant  2.24  0.6–8.5  0.34  0.23  1.62  0.4–6.1  0.51  0.48  Biologics  1.44  0.4–5.2  0.75  0.57  5.77  1.0–32.7  0.05  0.03  B. Other data                  Disease active  4.88  1.2–20.3  0.03  0.02  1.53  0.5–5.0  0.55  0.07  BMI > 30  0.28  0.1–1.0  0.06  0.14  0.88  0.2–4.3  1.00  0.73  Previous surgery  1.24  0.4–3.9  0.78  0.86  -  -  -  -  Breastfed  1.13  0.3–4.3  1.00  0.41  0.70  0.2–3.2  0.72  0.64  IBD, inflammatory bowel disease; 5-ASA, 5-aminosalicylic acid; BMI, body mass index. View Large To further investigate the relationship between BCoAT gene content and inflammation, faecal concentrations of calprotectin were compared in the aforementioned groups. Thereby, when CD patients were sub-grouped according to 9.5 log10 copies BCoAT/g cut-off point, average calprotectin level was higher for patients with CD and low BCoAT gene content [512 ± 1301 vs 112 ± 217 µg/g], although this was not statistically significant [p = 0.08]. We observed that patients with UC displayed no difference in average calprotectin when split by BCoAT gene levels [422 ± 803 vs 440 ± 1178 µg/g]. 3.3. Microbial diversity Comparison of microbiota alpha [intra-individual] diversity across the study populations showed that patients with CD and UC had significantly lower diversity compared with control individuals [Figure 2A and C]. When patients with CD and UC were divided according to BCoAT gene content, a lower diversity [p < 0.05 for chao1 index and 0.05 < p < 0.1 for Shannon index] was observed for CD patients with BCoAT levels below 9.5 log10 copies BCoAT/g [Figure 2B and D]. In contrast, this decrease was not found in patients with UC who displayed a similar level of alpha diversity independently of their BCoAT concentration. A positive correlation was observed between BCoAT gene content and alpha diversity, being strongest for patients with CD and low BCoAT levels [Supplementary Figure 2, available as Supplementary data at ECCO-JCC online]. Figure 2. View largeDownload slide Comparison of microbiota alpha diversity across groups. Tukey plots showing alpha diversity chao1 index for CD, UC and control groups [A] and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [B]. Tukey plots showing alpha diversity Shannon index for CD, UC and control groups [C]; and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [D]. Statistical differences between CD ≤ 9.5 vs UC ≤ 9.5, CD > 9.5 vs UC > 9.5, CD ≤ 9.5 vs UC > 9.5 and CD > 9.5 vs UC ≤ 9.5 are not shown. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Figure 2. View largeDownload slide Comparison of microbiota alpha diversity across groups. Tukey plots showing alpha diversity chao1 index for CD, UC and control groups [A] and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [B]. Tukey plots showing alpha diversity Shannon index for CD, UC and control groups [C]; and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [D]. Statistical differences between CD ≤ 9.5 vs UC ≤ 9.5, CD > 9.5 vs UC > 9.5, CD ≤ 9.5 vs UC > 9.5 and CD > 9.5 vs UC ≤ 9.5 are not shown. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Regarding microbiota beta [inter-individual] diversity, PCoA based on weighted UniFrac distances showed a shift in microbiota composition for patients with CD and UC away that of healthy controls, which was even more pronounced for CD [Figure 3A]. Categorising patients with CD and UC according to BCoAT gene content revealed that patients with CD and low BCoAT gene levels more often displayed a statistically significant shift along principal co-ordinate 1 [PC1] axis compared with healthy controls and UC patients than did those with CD and high BCoAT gene levels, with the latter showing a shift towards control individuals [Figure 3B]. Figure 3. View largeDownload slide Comparison of microbiota beta diversity across groups. Weighted UniFrac VST principal coordinate analysis [PCoA] and violin plots representing PCoA points along PC1 and PC2 axis for CD, UC, and control groups [A]; and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT /g cut-off point [B]. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Figure 3. View largeDownload slide Comparison of microbiota beta diversity across groups. Weighted UniFrac VST principal coordinate analysis [PCoA] and violin plots representing PCoA points along PC1 and PC2 axis for CD, UC, and control groups [A]; and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT /g cut-off point [B]. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. 3.4. Microbiota compositional changes in known butyrogenic bacteria We investigated whether the observed differences in microbiota composition were due in part to alterations in known butyrate producers, comprising two groups of bacteria [Clostridium clusters IV and XIVa], three genera [Anaerostipes, Butyricicoccus, and Roseburia] and four species [Blautia faecis, Eubacterium hallii, Faecalibacterium prausnitzii, and Ruminococcus torques]. We observed statistically significant changes for Clostridium cluster IV, Roseburia, and Faecalibacterium prausnitzii in the comparison between controls and patients with CD, but no butyrogenic taxa were altered between controls and patients with UC [Supplementary Table 2, available as Supplementary data at ECCO-JCC online]. Those changes were even more evident for patients with CD when considering BCoAT gene content [Supplementary Table 2; and Figure 4]. We found significant reductions for six of the taxa analysed between CD patients with low BCoAT concentration and controls. Furthermore, Butyricicoccus and Roseburia genera [and also Clostridium cluster XIVa and Eubacterium hallii with an adjusted p-value between 0.05 and 0.10] showed a decrease in CD patients with low BCoAT level compared with those with high BCoAT gene content. Accordingly, the only significant difference between CD patients with high BCoAT levels and controls was found for Clostridium cluster IV. In patients with UC, significant changes in bacterial composition between those with low BCoAT gene content and controls were limited to a decrease in Roseburia genus, while no changes were observed when comparing UC patients with high BCoAT levels and controls. Figure 4. View largeDownload slide Tukey boxplots for bacteria composition in healthy control samples and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point. Percentage of each of the group/genus/species considered with respect to total bacteria in stool samples are represented on the Y axis. Groups: Clostridium cluster IV and Clostridium cluster XIVa. Genera: Roseburia, Anaerostipes, and Butyricicoccus. Species: Faecalibacterium prausnitzii and Eubacterium hallii. Statistical differences between CD ≤ 9.5 vs UC ≤ 9.5, CD > 9.5 vs UC > 9.5, CD ≤ 9.5 vs UC > 9.5 and CD > 9.5 vs UC ≤ 9.5 are not shown. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Figure 4. View largeDownload slide Tukey boxplots for bacteria composition in healthy control samples and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point. Percentage of each of the group/genus/species considered with respect to total bacteria in stool samples are represented on the Y axis. Groups: Clostridium cluster IV and Clostridium cluster XIVa. Genera: Roseburia, Anaerostipes, and Butyricicoccus. Species: Faecalibacterium prausnitzii and Eubacterium hallii. Statistical differences between CD ≤ 9.5 vs UC ≤ 9.5, CD > 9.5 vs UC > 9.5, CD ≤ 9.5 vs UC > 9.5 and CD > 9.5 vs UC ≤ 9.5 are not shown. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. 3.5. Influence of diet on butyrate-synthetic capacity Differences in diet were also considered as a factor that could affect butyrate synthesis in the gut by modifying microbial composition [Table 3]. We observed a lower vegetable and fruit intake in patients with CD compared with healthy controls, and they showed increased intake of processed low fibre bread [white bread] and high sugar foods compared with controls. When patients with CD and UC were sub-grouped according to BCoAT gene content, patients with CD and high BCoAT gene concentration had a larger intake of nuts than those with low BCoAT levels, whereas no dietary changes were found for patients with UC. In the comparison of dietary habits among these groups and controls, we observed major significant differences between healthy controls and CD patients with low BCoAT gene content, with the latter showing reduced intake of certain foods containing fibre [Supplementary Table 3, available as Supplementary data at ECCO-JCC online] such as vegetables, fruits, cereals, brown/wholemeal bread, and nuts, and increased intake of high sugar food and white bread. Minor changes were noticed also in the diet of patients with UC and high BCoAT gene content compared with healthy controls. Table 3. Statistically significant changes observed for food categories in diet comparisons from data obtained through a food frequency questionnaire. Mean ± SD [standard deviation] refers to average number of times per month a food in each category was consumed. In the right column, p-values obtained in Wilcoxon test and adjusted for multiple testing are shown. Healthy controls vs CD  Healthy controls Mean ± SD  Patients with CD Mean ± SD  p-Value  Fruit  17.0 ± 17.0  11.0 ± 12.1  < 0.001  Vegetables  36.5 ± 38.3  18.5 ± 13.7  < 0.001  High sugar food  22.5 ± 17.9  29.7 ± 22.0  < 0.001  White bread  4.5 ± 8.1  8.5 ± 10.0  < 0.01  CD > 9.5 vs CD ≤ 9.5 copies BCoAT gene/g  CD > 9.5 copies BCoAT gene/g Mean ± SD  CD ≤ 9.5 copies BCoAT gene/g Mean ± SD  p-value  Nuts  3.9 ± 5.0  0.9 ± 1.9  < 0.01  Healthy controls vs CD ≤ 9.5 copies BCoAT gene/g  Healthy controls Mean ± SD  CD ≤ 9.5 copies BCoAT gene/g Mean ± SD  p-Value  Brown/wholemeal bread  8.5 ± 12.2  4.9 ± 8.2  < 0.05  Fruit  17.0 ± 17.0  7.9 ± 7.6  < 0.001  High fibre cereals  4.6 ± 6.0  2.2 ± 4.0  < 0.01  Nuts  3.7 ± 13.4  0.9 ± 1.9  < 0.05  Vegetables  36.5 ± 38.3  13.9 ± 9.2  < 0.001  High sugar food  22.5 ± 17.9  34.7 ± 20.9  < 0.01  White bread  4.5 ± 8.1  8.2 ± 10.9  < 0.05  Healthy controls vs UC > 9.5 copies BCoAT gene/g  Healthy controls Mean ± SD  UC > 9.5 copies BCoAT gene/g Mean ± SD  p-Value  Fruit  17.0 ± 17.0  11.3 ± 7.5  < 0.05  Vegetables  36.5 ± 38.3  26.6 ± 29.3  < 0.05  High sugar food  22.5 ± 17.9  40.1 ± 30.5  < 0.05  Healthy controls vs CD  Healthy controls Mean ± SD  Patients with CD Mean ± SD  p-Value  Fruit  17.0 ± 17.0  11.0 ± 12.1  < 0.001  Vegetables  36.5 ± 38.3  18.5 ± 13.7  < 0.001  High sugar food  22.5 ± 17.9  29.7 ± 22.0  < 0.001  White bread  4.5 ± 8.1  8.5 ± 10.0  < 0.01  CD > 9.5 vs CD ≤ 9.5 copies BCoAT gene/g  CD > 9.5 copies BCoAT gene/g Mean ± SD  CD ≤ 9.5 copies BCoAT gene/g Mean ± SD  p-value  Nuts  3.9 ± 5.0  0.9 ± 1.9  < 0.01  Healthy controls vs CD ≤ 9.5 copies BCoAT gene/g  Healthy controls Mean ± SD  CD ≤ 9.5 copies BCoAT gene/g Mean ± SD  p-Value  Brown/wholemeal bread  8.5 ± 12.2  4.9 ± 8.2  < 0.05  Fruit  17.0 ± 17.0  7.9 ± 7.6  < 0.001  High fibre cereals  4.6 ± 6.0  2.2 ± 4.0  < 0.01  Nuts  3.7 ± 13.4  0.9 ± 1.9  < 0.05  Vegetables  36.5 ± 38.3  13.9 ± 9.2  < 0.001  High sugar food  22.5 ± 17.9  34.7 ± 20.9  < 0.01  White bread  4.5 ± 8.1  8.2 ± 10.9  < 0.05  Healthy controls vs UC > 9.5 copies BCoAT gene/g  Healthy controls Mean ± SD  UC > 9.5 copies BCoAT gene/g Mean ± SD  p-Value  Fruit  17.0 ± 17.0  11.3 ± 7.5  < 0.05  Vegetables  36.5 ± 38.3  26.6 ± 29.3  < 0.05  High sugar food  22.5 ± 17.9  40.1 ± 30.5  < 0.05  CD, Crohn’s disease; UC, ulcerative colitis. View Large In the PCA analysis, we detected that the diet of patients with CD exhibited a pronounced shift away from the healthy controls’ diet, whereas the diet of patients with UC displayed a lesser shift but still statistically significant for PC2 axis [Figure 5A]. Interestingly, the diet of patients with CD and higher BCoAT gene content, but not for UC, exhibited a shift towards healthy controls’ diet in the PCA plot and showed lower statistical difference compared with controls [p < 0.05] than did patients with CD and low BCoAT gene levels [p < 0.001] in the PC2 axis [Figure 5B]. No major differences were observed either in the statistical analysis or in the PCA plot for the comparison between the 20 healthy controls with the highest BCoAT content and the 20 controls with the lowest BCoAT level [Figure 5C]. We investigated also if patients with CD showed differences in diet according to their disease location or behaviour, but we found no significant differences in the statistical analysis and no major shifts in the PCA plot [Supplementary Figure 3, available as Supplementary data at ECCO-JCC online]. Figure 5. View largeDownload slide Comparison of dietary data by principal component analysis [PCA] and by Wilcoxon tests [for PCA points along PC1 and PC2 axis, which were depicted as violin plots] for CD [n = 63], UC [n = 56], and control [n = 67] groups [A]; for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [CD ≤ 9.5 n = 27, CD > 9.5 n = 17, UC ≤ 9.5 n = 18, UC > 9.5 n = 27] [B]; and for the 20 healthy controls with the highest BCoAT gene content and the 20 healthy controls with the lowest BCoAT concentration [C]. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Figure 5. View largeDownload slide Comparison of dietary data by principal component analysis [PCA] and by Wilcoxon tests [for PCA points along PC1 and PC2 axis, which were depicted as violin plots] for CD [n = 63], UC [n = 56], and control [n = 67] groups [A]; for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [CD ≤ 9.5 n = 27, CD > 9.5 n = 17, UC ≤ 9.5 n = 18, UC > 9.5 n = 27] [B]; and for the 20 healthy controls with the highest BCoAT gene content and the 20 healthy controls with the lowest BCoAT concentration [C]. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Finally, procrustes analysis of the relationship between microbiota composition and diet revealed a significant correlation [p < 0.001], suggesting that diet had a direct or indirect effect on microbiota composition [Supplementary Figure 4, available as Supplementary data at ECCO-JCC online]. 4. Discussion Our results showed a reduction in the genetic capacity of colonic microbiota to produce butyrate, which is not only more pronounced in CD than in UC, but it also appears more relevant for CD activity, as low BCoAT gene content was found significantly associated only with active CD. Within CD patients, ileal location and stenotic behaviour showed the lowest BCoAT gene concentrations, being significantly decreased compared with colonic location and non-stenotic/non-penetrating behaviour, respectively. In accordance with this, the inflammation levels, disease-related changes in microbiota composition, and decreased percentage of butyrate-producers were greater in patients with CD having low BCoAT gene content. In addition, we noted dietary differences between healthy individuals and CD patients with low BCoAT gene content, indicating a decreased intake of fibre-rich food [like vegetables, fruits, or cereals] for the latter. Taken together, these results confirm existing knowledge about a decrease in butyrate-producer species in IBD, and contribute new evidence supporting the idea that a reduction in butyrate synthesis genetic capacity is more pronounced and relevant for CD than for UC, with diet being a plausible determinant behind butyrate decline. We observed a significant reduction in BCoAT gene levels for active and inactive patients with CD and active patients with UC as compared with controls. It constitutes a novel approach for evaluating butyrate alterations in IBD, since a dozen of previous studies were focused on faecal butyrate determination by metabolomic approaches. Although most of the studies showed a decrease in butyrate levels in patients with IBD, their results were discrepant. Some reported greater reductions for patients with CD than in UC,13 or only for active Crohn’s.11,14,33 Others described differences between patients with UC and healthy controls,14,34 but some reports were conflicting and found changes only for patients with active UC35 or did not observe any such differences.36 Although the concentration of butyrate in faecal samples may reflect the relative abundance of butyrate-producing bacteria in the gut, chromatographic measurements have limitations.37 First, they may be heavily influenced by butyrate uptake by the host. Second, physiological processes in the gut such as binding, degradation, and mucosal absorption, hamper the reliable estimation of butyrate concentration by direct analysis of faecal specimens. Furthermore, the diverse techniques employed [mainly HPLC, GC or GC/MS] coupled with different detector types and different extraction and derivatisation procedures confound standardisation. In contrast, qPCR is more standardised than the diverse chromatographic techniques. Therefore, we believe that this qPCR analysis of BCoAT gene content represents a reliable and useful method for evaluating butyrate synthesis capacity of gut microbes, which could complement chromatographic determinations. For example, a recent study showed significant differences in BCoAT gene content associated with a microbial shift between patients with UC responding or not responding to faecal microbiota transplantation, whereas no changes were observed in faecal butyrate levels by HPLC.38 Some differences in baseline characteristics were found between healthy control volunteers and our IBD cohort for age, obesity, breastfeeding, and caesarean birth. We have analysed these variables to check whether they could have an effect on BCoAT gene content. Age was not correlated with BCoAT levels and obesity and breastfeeding showed no association with BCoAT subgroups in both univariate and multivariate logistic regression analysis. In addition, the higher percentage of overweight individuals in our IBD cohort compared with healthy controls seems to be in agreement with recent epidemiological data.39 Birth by caesarean section was not considered, due to low numbers in each group for comparison, but it was described that it only affects microbiota during the first 6 months of life.40 Our data indicated a significant decrease in microbiota alpha diversity and a relevant shift in beta diversity away from healthy controls for patients with CD and UC, in agreement with previous reports.15 Both changes in diversity were more prominent for patients with CD and, only in them we observed a higher microbiota alpha diversity and a more similar microbiota to that of controls in patients with high BCoAT gene concentration. Our analysis of changes in particular butyrogenic taxa also proved that such alterations were smaller in patients with CD and high BCoAT gene content. Within the Firmicutes phylum, two 16S-rRNA-defined phylogenetics groups, Clostridium cluster IV and XIVa, include the majority of the bacteria identified as capable of synthesising butyrate. A previous study of butyrogenic bacteria diversity in human faecal samples using BCoAT amplification revealed the importance of the Roseburia group and E. hallii, with an unexpected under-representation for F. prausnitzii which could be explained by a selective loss of F. prausnitzii BCoAT sequences during the cloning step.41 Strain SS2-1 that was reported in this work as a major butyrate-producer using the BCoAT enzyme, was later identified as Anaerostipes hadrus.42 All these taxa were selected for comparison [relative abundance] in the stool microbiota, jointly with genus Butyricicoccus20 and B. faecis and R. torques species.43 We observed reduced relative abundances for six taxa in patients with CD and low BCoAT content compared with healthy controls, whereas solely one taxon was decreased for the same BCoAT sub-group of patients with UC. As only one taxon, Clostridium cluster IV, was declined for patients with CD and high BCoAT gene concentration when compared with controls, increased BCoAT levels correlated with a relevant improvement in butyrogenic bacteria abundance. Our findings for butyrate-producers in patients with CD are in general agreement with previous reports on this topic.44 Several studies, primarily in UC, described promising but inconclusive results in ameliorating inflammation and symptoms with therapeutic administration of butyrate.10 As our results showed no difference in BCoAT gene content between patients with inactive UC and healthy controls, it could be hypothesised that interventions aimed to increase butyrate levels in their colon would result in minimal effects. A previous interventional study confirmed this hypothesis, since administration of rectal butyrate enemas for 3 weeks in patients with inactive UC caused minor improvements on inflammatory and oxidative stress measurements.45 Butyrate was reported as possibly effective in achieving CD remission when administered orally.46,47 The alterations in the microbiota and the reduction of known butyrate producers in patients with CD, which were higher in those with low BCoAT gene content, suggest that microbiota modulation, either directly or promoted through prebiotics, may provide more consistent and efficient changes in butyrate concentrations in the gut compared with direct butyrate administration. Prebiotic compounds described as stimulators of butyrate synthesis are resistant starch, oat bran, sorbitol, galacturonic and glucuronic acid, inulin, and fructo-oligosaccharides.48 Regarding probiotics, promising results for butyrate producers that could be used as pharmabiotics by themselves to attenuate IBD were obtained for F. prausnitzii, and its culture supernatant,49–52 and B. paellicorum,20,53 although negative effects were also reported for the inoculation of the butyrate producer strain Anaerostipes hadrus BPB5 in a DSS-induced colitis mouse model.54 Another means to achieve more long-term changes in the gut microbiota and thus in butyrate levels might be diet. The impact of diet on butyrate and other SCFA produced by the microbiota has been widely investigated, in both epidemiological and interventional studies.55 For example, European children had almost four times lower levels of butyrate than children living in a rural Burkina Faso village, suggesting that a diet rich in fibre and low in sugar and fat could favour SCFA-producing bacteria.56 Therefore, it is not surprising that dietary factors related to SCFA could affect the course and treatment of IBD.57 Our procrustes analysis showed that diet had a noticeable effect on the microbiota. In particular, patients with CD and low BCoAT gene content had a more ‘westernised’ diet with lower intake of fruit, high fibre cereals, nuts, brown/wholemeal bread, and vegetables compared with controls, and also a higher intake of high sugar foods and white bread, which are sources of carbohydrates. Moreover, the food group which includes nuts, which tended to be consumed to a greater degree by CD patients with higher BCoAT levels than by those with lower levels of BCoAT, may represent an alternative source of fibre to increase butyrate. Of course, recommendations regarding dietary fibre must be judicious particularly in patients at risk of stricture formation, where reduced dietary fibre is a traditional recommendation. There were some limitations in our study. First, although disease activity was not confirmed by endoscopy, a notional gold standard, we supplemented the physicians’ assessment with conventional clinical indices and faecal calprotectin. Second, the primers employed were designed based on a limited number of sequences, but they were proved to capture the known butyrate producers with higher compositional representation in the gut microbiota.41 Third, our analysis of butyrate-producing capacity was limited to BCoAT gene content, although other factors are known to affect butyrate availability and utilisation; for example, butyrate uptake/transportation by colonocytes may be impaired in IBD.58 In addition, butyrate synthesis through the butyrate kinase pathway or from other alternative sources, such as lysine or 4-aminobutyrate/succinate, were not considered, but these pathways appear to be much less important from a quantitative point of view.59 Regarding dietary data, the FFQ employed did not provide information of portion sizes, which did not allow calculation of dietary intake for nutrients. It is beyond the aim of the present study evaluate whether reduced butyrate synthesis capacity is cause or consequence of the disease, even though we hypothesise that normalisation of the gut microbiota butyrogenic capacity may associate with an improvement in patients with CD. New studies will ultimately be required to investigate whether BCoAT gene content determination by qPCR could be helpful for patient management in IBD. In conclusion, our study confirms evidence for butyrate alterations in IBD and extends current knowledge by showing that a decrease in the genetic capacity of the gut microbiota to synthesise butyrate is associated with active disease, higher degree of inflammation, and greater changes in the microbiota in patients with CD but not in those with UC. Since dietary habits influence the butyrate synthetic capacity of the microbiota, the judicious use of dietary fibre deserves reconsideration as a therapeutic strategy. Funding This work was supported by Science Foundation Ireland [SFI] under Grant Numbers 11/SIRG/B2162 and SFI/12/RC/2273. EJLM conducted part of this research under a contract funded mostly with a research grant from the European Crohn’s and Colitis Organization [year 2014]. Conflict of Interest APC Microbiome Institute receives funding from several private companies, but none of them has been involved in the present study. Author Contributions EJLM, AGC, MJC: study concept and data analysis; EJLM, AGC, FS, MJC: output formatting and manuscript writing; EJLM: sample processing and laboratory work; EJLM, JFCG: performing calprotectin assay; EJLM, AGC, CM, DS, CGMG, SAJ, FS, MJC: patient recruitment and data collection; JN, CH: recruitment of healthy controls; all authors: final review of manuscript. All authors have approved the final version of the manuscript. Supplementary Data Supplementary data are available at ECCO-JCC online. Acknowledgments We would like to thank Dr Feargal Ryan, Ms Emily Power, Mr Alvaro Lopez-Gallardo, and Dr Kalaimathi Govindarajan [University College Cork] for helping in metadata processing, Ms Ángeles Cabezas-Martínez and Mrs María Jesús Rocha-Bogas [Complejo Hospitalario de Toledo] for assisting in calprotectin assay, Dr Fabien Cousin and Dr Guillaume Borrel [former post-doctoral researchers at University College Cork] for helping in setting up qPCR assay, Dr Janette Walton [School of Food and Nutritional Science, University College Cork] for collaborating in dietary data interpretation, and Ms Margot Hurley and Ms Catherine O’Riordan [Cork University Hospital] for collecting stool samples and providing questionnaires. References 1. Sheehan D, Moran C, Shanahan F. The microbiota in inflammatory bowel disease. J Gastroenterol  2015; 50: 495– 507. Google Scholar CrossRef Search ADS PubMed  2. Louis P, Duncan SH, McCrae SI, Millar J, Jackson MS, Flint HJ. Restricted distribution of the butyrate kinase pathway among butyrate-producing bacteria from the human colon. J Bacteriol  2004; 186: 2099– 106. Google Scholar CrossRef Search ADS PubMed  3. Hamer HM, Jonkers D, Venema K, Vanhoutvin S, Troost FJ, Brummer RJ. Review article: the role of butyrate on colonic function. Aliment Pharmacol Ther  2008; 27: 104– 19. Google Scholar CrossRef Search ADS PubMed  4. Plöger S, Stumpff F, Penner GBet al.   Microbial butyrate and its role for barrier function in the gastrointestinal tract. Ann N Y Acad Sci  2012; 1258: 52– 9. Google Scholar CrossRef Search ADS PubMed  5. Zhang LS, Davies SS. Microbial metabolism of dietary components to bioactive metabolites: opportunities for new therapeutic interventions. Genome Med  2016; 8: 46. Google Scholar CrossRef Search ADS PubMed  6. Segain JP, Raingeard de la Blétière D, Bourreille Aet al.   Butyrate inhibits inflammatory responses through NFkappaB inhibition: implications for Crohn’s disease. Gut  2000; 47: 397– 403. Google Scholar CrossRef Search ADS PubMed  7. Arpaia N, Campbell C, Fan Xet al.   Metabolites produced by commensal bacteria promote peripheral regulatory T-cell generation. Nature  2013; 504: 451– 5. Google Scholar CrossRef Search ADS PubMed  8. Furusawa Y, Obata Y, Fukuda Set al.   Commensal microbe-derived butyrate induces the differentiation of colonic regulatory T cells. Nature  2013; 504: 446– 50. Google Scholar CrossRef Search ADS PubMed  9. Smith PM, Howitt MR, Panikov Net al.   The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis. Science  2013; 341: 569– 73. Google Scholar CrossRef Search ADS PubMed  10. Sossai P. Butyric acid: what is the future for this old substance? Swiss Med Wkly  2012; 142: w13596. Google Scholar PubMed  11. De Preter V, Joossens M, Ballet Vet al.   Metabolic profiling of the impact of oligofructose-enriched inulin in Crohn’s disease patients: a double-blinded randomized controlled trial. Clin Transl Gastroenterol  2013; 4: e30. Google Scholar CrossRef Search ADS PubMed  12. De Preter V, Machiels K, Joossens Met al.   Faecal metabolite profiling identifies medium-chain fatty acids as discriminating compounds in IBD. Gut  2015; 64: 447– 58. Google Scholar CrossRef Search ADS PubMed  13. Marchesi JR, Holmes E, Khan Fet al.   Rapid and noninvasive metabonomic characterization of inflammatory bowel disease. J Proteome Res  2007; 6: 546– 51. Google Scholar CrossRef Search ADS PubMed  14. Takaishi H, Matsuki T, Nakazawa Aet al.   Imbalance in intestinal microflora constitution could be involved in the pathogenesis of inflammatory bowel disease. Int J Med Microbiol  2008; 298: 463– 72. Google Scholar CrossRef Search ADS PubMed  15. Kostic AD, Xavier RJ, Gevers D. The microbiome in inflammatory bowel disease: current status and the future ahead. Gastroenterology  2014; 146: 1489– 99. Google Scholar CrossRef Search ADS PubMed  16. Frank DN, St Amand AL, Feldman RA, Boedeker EC, Harpaz N, Pace NR. Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases. Proc Natl Acad Sci U S A  2007; 104: 13780– 5. Google Scholar CrossRef Search ADS PubMed  17. Cao Y, Shen J, Ran ZH. Association between Faecalibacterium prausnitzii reduction and inflammatory bowel disease: a meta-analysis and systematic review of the literature. Gastroenterol Res Pract  2014; 2014: 872725. Google Scholar PubMed  18. Rajilić-Stojanović M, Shanahan F, Guarner F, de Vos WM. Phylogenetic analysis of dysbiosis in ulcerative colitis during remission. Inflamm Bowel Dis  2013; 19: 481– 8. Google Scholar CrossRef Search ADS PubMed  19. Gevers D, Kugathasan S, Denson LAet al.   The treatment-naive microbiome in new-onset Crohn’s disease. Cell Host Microbe  2014; 15: 382– 92. Google Scholar CrossRef Search ADS PubMed  20. Eeckhaut V, Machiels K, Perrier Cet al.   Butyricicoccus pullicaecorum in inflammatory bowel disease. Gut  2013; 62: 1745– 52. Google Scholar CrossRef Search ADS PubMed  21. Morgan XC, Tickle TL, Sokol Het al.   Dysfunction of the intestinal microbiome in inflammatory bowel disease and treatment. Genome Biol  2012; 13: R79. Google Scholar CrossRef Search ADS PubMed  22. Erickson AR, Cantarel BL, Lamendella Ret al.   Integrated metagenomics/metaproteomics reveals human host-microbiota signatures of Crohn’s disease. PLoS One  2012; 7: e49138. Google Scholar CrossRef Search ADS PubMed  23. Claesson MJ, Jeffery IB, Conde Set al.   Gut microbiota composition correlates with diet and health in the elderly. Nature  2012; 488: 178– 84. Google Scholar CrossRef Search ADS PubMed  24. Harvey RF, Bradshaw JM. A simple index of Crohn’s-disease activity. Lancet  1980; 1: 514. Google Scholar CrossRef Search ADS PubMed  25. Powell-Tuck J, Day DW, Buckell NA, Wadsworth J, Lennard-Jones JE. Correlations between defined sigmoidoscopic appearances and other measures of disease activity in ulcerative colitis. Dig Dis Sci  1982; 27: 533– 7. Google Scholar CrossRef Search ADS PubMed  26. Walmsley RS, Ayres RC, Pounder RE, Allan RN. A simple clinical colitis activity index. Gut  1998; 43: 29– 32. Google Scholar CrossRef Search ADS PubMed  27. Clooney AG, Bernstein CN, Leslie WDet al.   A comparison of the gut microbiome between long-term users and non-users of proton pump inhibitors. Aliment Pharmacol Ther  2016; 43: 974– 84. Google Scholar CrossRef Search ADS PubMed  28. Klindworth A, Pruesse E, Schweer Tet al.   Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res  2013; 41: e1. Google Scholar CrossRef Search ADS PubMed  29. Louis P, Flint HJ. Development of a semiquantitative degenerate real-time pcr-based assay for estimation of numbers of butyryl-coenzyme A [CoA] CoA transferase genes in complex bacterial samples. Appl Environ Microbiol  2007; 73: 2009– 12. Google Scholar CrossRef Search ADS PubMed  30. Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics  2011; 27: 2194– 200. Google Scholar CrossRef Search ADS PubMed  31. Schloss PD, Westcott SL, Ryabin Tet al.   Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol  2009; 75: 7537– 41. Google Scholar CrossRef Search ADS PubMed  32. Allard G, Ryan FJ, Jeffery IB, Claesson MJ. SPINGO: a rapid species-classifier for microbial amplicon sequences. BMC Bioinformatics  2015; 16: 324. Google Scholar CrossRef Search ADS PubMed  33. Bjerrum JT, Wang Y, Hao Fet al.   Metabonomics of human fecal extracts characterize ulcerative colitis, Crohn’s disease and healthy individuals. Metabolomics  2015; 11: 122– 33. Google Scholar CrossRef Search ADS PubMed  34. Huda-Faujan N, Abdulamir AS, Fatimah ABet al.   The impact of the level of the intestinal short chain fatty acids in inflammatory bowel disease patients versus healthy subjects. Open Biochem J  2010; 4: 53– 8. Google Scholar CrossRef Search ADS PubMed  35. Kumari R, Ahuja V, Paul J. Fluctuations in butyrate-producing bacteria in ulcerative colitis patients of North India. World J Gastroenterol  2013; 19: 3404– 14. Google Scholar CrossRef Search ADS PubMed  36. Machiels K, Joossens M, Sabino Jet al.   A decrease of the butyrate-producing species Roseburia hominis and Faecalibacterium prausnitzii defines dysbiosis in patients with ulcerative colitis. Gut  2014; 63: 1275– 83. Google Scholar CrossRef Search ADS PubMed  37. Verbeke KA, Boobis AR, Chiodini Aet al.   Towards microbial fermentation metabolites as markers for health benefits of prebiotics. Nutr Res Rev  2015; 28: 42– 66. Google Scholar CrossRef Search ADS PubMed  38. Fuentes S, Rossen NG, van der Spek MJet al.   Microbial shifts and signatures of long-term remission in ulcerative colitis after faecal microbiota transplantation. ISME J  2017; 11: 1877– 89. Google Scholar CrossRef Search ADS PubMed  39. Moran C, Sheehan D, Shanahan F. The changing phenotype of inflammatory bowel disease. Gastroenterol Res Pract  2016; 2016: 1619053. Google Scholar CrossRef Search ADS PubMed  40. Rutayisire E, Huang K, Liu Y, Tao F. The mode of delivery affects the diversity and colonization pattern of the gut microbiota during the first year of infants’ life: a systematic review. BMC Gastroenterol  2016; 16: 86. Google Scholar CrossRef Search ADS PubMed  41. Louis P, Young P, Holtrop G, Flint HJ. Diversity of human colonic butyrate-producing bacteria revealed by analysis of the butyryl-CoA:acetate CoA-transferase gene. Environ Microbiol  2010; 12: 304– 14. Google Scholar CrossRef Search ADS PubMed  42. Allen-Vercoe E, Daigneault M, White Aet al.   Anaerostipes hadrus comb. nov., a dominant species within the human colonic microbiota; reclassification of Eubacterium hadrum Moore et al. 1976. Anaerobe  2012; 18: 523– 9. Google Scholar CrossRef Search ADS PubMed  43. Takahashi K, Nishida A, Fujimoto Tet al.   Reduced abundance of butyrate-producing bacteria species in the fecal microbial community in Crohn’s Disease. Digestion  2016; 93: 59– 65. Google Scholar CrossRef Search ADS PubMed  44. Wright EK, Kamm MA, Teo SM, Inouye M, Wagner J, Kirkwood CD. Recent advances in characterizing the gastrointestinal microbiome in Crohn’s disease: a systematic review. Inflamm Bowel Dis  2015; 21: 1219– 28. Google Scholar CrossRef Search ADS PubMed  45. Hamer HM, Jonkers DM, Vanhoutvin SAet al.   Effect of butyrate enemas on inflammation and antioxidant status in the colonic mucosa of patients with ulcerative colitis in remission. Clin Nutr  2010; 29: 738– 44. Google Scholar CrossRef Search ADS PubMed  46. Di Sabatino A, Morera R, Ciccocioppo Ret al.   Oral butyrate for mildly to moderately active Crohn’s disease. Aliment Pharmacol Ther  2005; 22: 789– 94. Google Scholar CrossRef Search ADS PubMed  47. Russo I, Luciani A, De Cicco P, Troncone E, Ciacci C. Butyrate attenuates lipopolysaccharide-induced inflammation in intestinal cells and Crohn’s mucosa through modulation of antioxidant defense machinery. PLoS One  2012; 7: e32841. Google Scholar CrossRef Search ADS PubMed  48. Li RW, Li CJ. Enhancing butyrate biosynthesis in the gut for health benefits. In: Butyrate: Food Sources, Functions and Health Benefits. Hauppauge , NY: USA Nova Science Publishers; 2014. 49. Miquel S, Martín R, Rossi Oet al.   Faecalibacterium prausnitzii and human intestinal health. Curr Opin Microbiol  2013; 16: 255– 61. Google Scholar CrossRef Search ADS PubMed  50. Qiu X, Zhang M, Yang X, Hong N, Yu C. Faecalibacterium prausnitzii upregulates regulatory T cells and anti-inflammatory cytokines in treating TNBS-induced colitis. J Crohns Colitis  2013; 7: e558– 68. Google Scholar CrossRef Search ADS PubMed  51. Martín R, Chain F, Miquel Set al.   The commensal bacterium Faecalibacterium prausnitzii is protective in DNBS-induced chronic moderate and severe colitis models. Inflamm Bowel Dis  2014; 20: 417– 30. Google Scholar CrossRef Search ADS PubMed  52. Sokol H, Pigneur B, Watterlot Let al.   Faecalibacterium prausnitzii is an anti-inflammatory commensal bacterium identified by gut microbiota analysis of Crohn disease patients. Proc Natl Acad Sci U S A  2008; 105: 16731– 6. Google Scholar CrossRef Search ADS PubMed  53. Eeckhaut V, Ducatelle R, Sas B, Vermeire S, Van Immerseel F. Progress towards butyrate-producing pharmabiotics: Butyricicoccus pullicaecorum capsule and efficacy in TNBS models in comparison with therapeutics. Gut  2014; 63: 367. Google Scholar CrossRef Search ADS PubMed  54. Zhang Q, Wu Y, Wang Jet al.   Accelerated dysbiosis of gut microbiota during aggravation of DSS-induced colitis by a butyrate-producing bacterium. Sci Rep  2016; 6: 27572. Google Scholar CrossRef Search ADS PubMed  55. Ríos-Covián D, Ruas-Madiedo P, Margolles A, Gueimonde M, de Los Reyes-Gavilán CG, Salazar N. Intestinal short chain fatty acids and their link with diet and human health. Front Microbiol  2016; 7: 185. Google Scholar CrossRef Search ADS PubMed  56. De Filippo C, Cavalieri D, Di Paola Met al.   Impact of diet in shaping gut microbiota revealed by a comparative study in children from Europe and rural Africa. Proc Natl Acad Sci U S A  2010; 107: 14691– 6. Google Scholar CrossRef Search ADS PubMed  57. Lee D, Albenberg L, Compher Cet al.   Diet in the pathogenesis and treatment of inflammatory bowel diseases. Gastroenterology  2015; 148: 1087– 106. Google Scholar CrossRef Search ADS PubMed  58. Thibault R, Blachier F, Darcy-Vrillon B, de Coppet P, Bourreille A, Segain JP. Butyrate utilization by the colonic mucosa in inflammatory bowel diseases: a transport deficiency. Inflamm Bowel Dis  2010; 16: 684– 95. Google Scholar CrossRef Search ADS PubMed  59. Vital M, Howe AC, Tiedje JM. Revealing the bacterial butyrate synthesis pathways by analyzing [meta]genomic data. MBio  2014; 5: e00889. Google Scholar CrossRef Search ADS PubMed  Copyright © 2017 European Crohn’s and Colitis Organisation (ECCO). Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Crohn's and Colitis Oxford University Press

Determinants of Reduced Genetic Capacity for Butyrate Synthesis by the Gut Microbiome in Crohn’s Disease and Ulcerative Colitis

Loading next page...
 
/lp/ou_press/determinants-of-reduced-genetic-capacity-for-butyrate-synthesis-by-the-yQAPE329Sx
Publisher
Elsevier Science
Copyright
Copyright © 2017 European Crohn’s and Colitis Organisation (ECCO). Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com
ISSN
1873-9946
eISSN
1876-4479
D.O.I.
10.1093/ecco-jcc/jjx137
Publisher site
See Article on Publisher Site

Abstract

Abstract Background and Aims Alterations in short chain fatty acid metabolism, particularly butyrate, have been reported in inflammatory bowel disease, but results have been conflicting because of small study numbers and failure to distinguish disease type, activity or other variables such as diet. We performed a comparative assessment of the capacity of the microbiota for butyrate synthesis, by quantifying butyryl-CoA:acetate CoA-transferase [BCoAT] gene content in stool from patients with Crohn’s disease [CD; n = 71], ulcerative colitis [UC; n = 58] and controls [n = 75], and determined whether it was related to active vs inactive inflammation, microbial diversity, and composition and/or dietary habits. Methods BCoAT gene content was quantified by quantitative polymerase chain reaction [qPCR]. Disease activity was assessed clinically and faecal calprotectin concentration measured. Microbial composition was determined by sequencing 16S rRNA gene. Dietary data were collected using an established food frequency questionnaire. Results Reduced butyrate-synthetic capacity was found in patients with active and inactive CD [p < 0.001 and p < 0.01, respectively], but only in active UC [p < 0.05]. In CD, low BCoAT gene content was associated with ileal location, stenotic behaviour, increased inflammation, lower microbial diversity, greater microbiota compositional change, and decreased butyrogenic taxa. Reduced BCoAT gene content in patients with CD was linked with a different regimen characterised by lower dietary fibre. Conclusions Reduced butyrate-synthetic capacity of the microbiota is more evident in CD than UC and may relate to reduced fibre intake. The results suggest that simple replacement of butyrate per se may be therapeutically inadequate, whereas manipulation of microbial synthesis, perhaps by dietary means, may be more appropriate. Butyrate synthesis, inflammatory bowel disease, microbiota 1. Introduction The inflammatory bowel diseases [IBD], Crohn’s disease [CD] and ulcerative colitis [UC], are heterogeneous and multifactorial disorders involving variable influences including genetic predisposition, immune-mediated tissue damage, and environmental modifiers, of which the gut microbiota may be the proximate environmental influence.1 Prominent among the microbial metabolites that interact with the host are short chain fatty acids [SCFA]. These are normally produced by the colonic microbiota by fermentation of complex undigested polysaccharides such as dietary fibres. The SCFA that has been most plausibly linked with IBD is butyrate, which can be synthesised from butyryl-CoA by two different enzymes: butyrate kinase and butyryl-CoA:acetate CoA-transferase [BCoAT], the latter being predominant in the human colonic ecosystem.2 Butyrate provides energy to the colonocytes, promoting epithelial cell growth. It also acts as a signalling molecule regulating cell proliferation and differentiation. Butyrate has been described to possess anti-carcinogenic, anti-lipogenic and anti-inflammatory potential, and may play a role in satiety, insulin sensitivity, oxidative stress, motility, sodium and water absorption, stimulation of mucous secretion, and increase of vascular flow.3–5 Butyrate inhibits inflammatory responses by decreasing proinflammatory cytokine expression via inhibition of NF-κB transcription factor activation in immune cells,6 and is considered a potent communicator to the immune system, acting mainly as an inhibitor of histone deacetylases. In that way, it favours efficient generation of regulatory T cells7,8 and homeostasis.9 Therefore, butyrate has been employed in several randomised clinical trials and interventional studies to analyse its efficacy for symptom relief, especially in diseases involving inflammation.10 Disturbed butyrate metabolism has been linked with IBD, but results have been either discrepant or conflicting in relation to CD vs UC, predominantly based on faecal levels, and seldom related to disease activity.11–14 Diminished colonic butyrogenic bacteria in IBD have been reported in faecal and mucosal samples from patients with IBD15 and most notably include Faecalibacterium prausnitzii16,17 and the Roseburia genus,18 which were also decreased in new-onset CD.19 In addition, lower numbers of the genus Butyricicoccus were reported in stools of patients with IBD.20 Alterations in butyrate production in IBD have been also suggested by metagenomic analysis.21,22 We hypothesised that disturbances in butyrogenic capacity of colonic microbiota may be more informative than faecal butyrate levels. However, direct quantification of the capacity of the gut microbiota to produce butyrate across active and inactive stages has not been performed. We have addressed this by quantifying BCoAT gene content in faecal microbiota from patients with CD and UC in comparison with controls, and results were related to clinical variables including disease activity and habitual diet. 2. Material and Methods 2.1. Study population Patients with CD and UC were recruited from dedicated IBD specialty clinics at Cork University Hospital and Mercy University Hospital, Cork, Ireland. Eligible patients were aged 18–67 years with a well-established diagnosis of chronic recurrent CD or UC. The following exclusion criteria were applied in patient recruitment: pregnant female; a history of total colectomy; significant acute or chronic coexisting illness; treatment with experimental drugs; and malignant or any concomitant end-stage organ disease. Healthy volunteers were recruited from the same geographical region and mostly among University College Cork staff. Informed consent was obtained from all study subjects and the study was approved by the Clinical Research Ethics Committee of the Cork Teaching Hospitals, University College Cork, National University of Ireland. All IBD subjects completed a clinical interview with a gastroenterologist and surveys detailing their demographic [age, gender, nationality, body mass index, smoking, breastfeeding, and born by caesarean section] and clinical characteristics [history of IBD-related surgeries, frequency of symptoms, IBD clinical indices, current medication, perceived stress, and history of other diseases and surgeries aside from IBD]. Patients with active IBD were encouraged to be enrolled so as to achieve a wide representation of disease activities in the studied cohort, with some of them having a change to a more aggressive treatment when recruited. Demographic characteristics were also obtained for control individuals. In addition, all subjects [except for eight patients with CD, two patients with UC, and eight control individuals] completed a detailed food frequency questionnaire [FFQ], as previously described by Claesson et al.23 Subjects were instructed to store faecal samples in a refrigerator from which they were transported within 24 h to the research laboratory and stored at 4°C. Samples were subdivided into aliquots of approximately 0.2 g and 0.5 g within 24 h of arrival for further DNA extraction and calprotectin measurement, respectively, and then frozen at -80°C. Wet weight of faecal samples was measured in those aliquots intended to be for DNA extraction before being frozen. Diarrhoeal samples with liquid consistency were annotated as such. 2.2. 16S rRNA gene sequencing and quantitative polymerase chain reaction assay A total number of 204 stool samples were included in this study [71 CD and 58 UC patients, and 75 healthy controls]. Some of the samples from diseased individuals were classified into two sub-groups [inactive and active] when fulfilling specific criteria: CD inactive when Harvey-Bradshaw Index [HBI]24 ≤ 4 and symptom activity described as ‘occasional [a day or more once a month]’, or ‘rare’, or ‘absent’; CD active when HBI ≥ 5 and symptom activity categorised as ‘active all the time’ or ‘active often [some days in each week]’; UC inactive when Powell-Tuck Index [PTI]25 ≤ 3 and Ulcerative Colitis Activity Index [UCAI]26 ≤ 4 and symptom activity described as ‘occasional’, or ‘rare’ or ‘absent’; and UC active when PTI ≥ 4 and UCAI ≥ 5 and symptom activity categorised as ‘active all the time’ or ‘active often’. According to these criteria, the number of patients in each sub-group was 16 CD active, 34 CD inactive, 19 UC active and 27 UC inactive. DNA extraction on faecal samples was carried out as previously described in Clooney et al.,27 using the QIAamp Fast DNA stool kit [Qiagen GmbH, Hilden, Germany] including repeated bead-beating. Library preparation of 16S rRNA gene amplicon for sequencing was performed following the Illumina [San Diego, CA, USA] recommendations with some modifications described in the aforementioned study.27 Primers for amplification of the V3-V4 hypervariable region of the 16S rRNA gene were selected from Klindworth et al.28 MiSeq sequencing of the 16S library was made at Eurofins Genomics [Ebersberg, Germany]. The number of copies of BCoAT gene was determined in genomic DNA [gDNA] samples [previously diluted 1:5] by quantitative polymerase chain reaction [qPCR]. Degenerate primers designed from the BCoAT sequence in four butyrate-producer species,29 BCoATscrF [GCIGAICATTTCACITGGAAYWSITGGCAYATG] and BCoATscrR [CCTGCCTTTGCAATRTCIACRAANGC] [Eurofins Genomics], were employed for amplification. The qPCR reaction was performed in a total volume of 20 µL containing LightCycler SYBR Green I Master [Roche, Penzberg, Germany], nuclease-free water, primers at a final concentration of 2 µM, and 2 µL of gDNA. Amplification was carried out in a LightCycler 480 Instrument II [Roche] under the following conditions: 95°C for 10 min, 40 cycles of 95°C for 10 s, 53°C for 15 s, and 72°C for 30 s with data acquisition at 72°C, and a stepwise increase of the temperature to 95°C [at 0.29°C/s] with continuous acquisition to obtain melting curve data. All samples were run in duplicates and mean values were used in calculations. To generate standard curves, calculated amounts of the linearised plasmid pGEM T Easy [Promega, Madison, WI, USA], in which the amplified region from the BCoAT gene had been inserted, were used. Plasmid concentration was measured using Nano-Drop 2000 Spectrophotometer [Thermo Scientific, Wilmington, DE, USA], and the number of copies was calculated using the DNA Copy Number and Dilution Calculator tool at Life Technologies website. Serial dilutions [from 107 to 10] were amplified in duplicates to create the standard curve for extrapolation of BCoAT gene copies using the Ct values of the samples. Results were multiplied according to the dilution factor, normalised as copies per gram of wet weight of faeces and, finally, expressed as log10 values. No diarrhoeal samples presenting liquid consistency were included in the study to avoid false reductions of bacterial load due to increased percentage of faecal water. Repeatability conditions were tested in a sample extracted six different times and processed in different qPCR assays. Coefficient of variation for these six samples was 1.5%, showing a good reproducibility between different DNA extractions and different qPCR runs. 2.3. Bioinformatics analysis A mean of 19464 (± 4869 standard deviation [SD]) reads per sample was obtained. Adaptor sequences were removed from V3-V4 16S rRNA gene amplicon reads using cutadapt with a 0.2 error rate allowance. The script fastq_merge [USEARCH version 7.0.1090] was employed to merge forward and reverse paired reads. Demultiplexing was carried out using split_libraries [QIIME] and only reads in the range of 390 bp to 465 bp in length and with an average PHRED score of Q25 and above were retained for downstream analysis. Operational taxonomic units [OTUs] were clustered using a de novo approach [QIIME] and removal of chimeric sequences was carried out using uchime_ref in conjunction with the ChimeraSlayer GOLD database.30 The OTU table was normalised using the variance stabilising technique [VST] from the DESeq R library. Taxonomy [phylum to genus] was assigned to OTUs using the Mothur implementation of the Ribosomal Database Project [RDP] classifier [version 11.4].31 Any sequence with less than an 80% bootstrap value was assigned as unclassified at that particular rank. SPINGO, with default parameters, was used for species and Clostridium cluster assignment on demultiplexed and quality filtered reads.32 2.4. Faecal calprotectin assay A -80°C frozen aliquot for each sample, containing approximately 0.5 g of faecal material, was used to measure calprotectin concentration with EliATM Calprotectin Immunoassay v2 [Thermo Scientific, Uppsala, Sweden], following manufacturer’s instructions. First, calprotectin was extracted from faeces using EliA Stool Extraction Kit. Samples were picked employing a rod with four notches that indicated the stool quantity required, and were introduced in the extraction buffer. Subsequently, they were vortexed and centrifuged. Finally, the resulting supernatant was processed in an ImmunoCap 250 autoanalyser [Thermo Scientific] set up with the buffers, conjugates, and reagents for EliATM calprotectin assay. A control curve was measured for each calprotectin run as a quality control. Results were expressed as µg/g faeces. Results below detection limit [< 3.8 µg/g] were accounted as 4 µg/g in calculations. Results above lineal range limit [> 6000 µg/g] were considered as 6000 µg/g in calculations. 2.5. Statistical analysis Normality distribution was assessed using the Kolmogorov-Smirnov test [when n ≥ 50] or the Shapiro-Wilk test [when n < 50]. Logarithmic transformation was applied to calprotectin when normality was not assumed. In cases where normality could not be assumed, non-parametric tests were employed [Kruskal-Wallis for multiple comparison and Mann-Whitney for independent comparisons]. Alpha and beta diversity were calculated using the phyloseq library, followed by Principal Coordinate Analysis [PCoA] graphed via the ggplot2 library. Wilcoxon rank sum tests were used to compare alpha and beta diversity between groups and taxa [after normalisation via proportions]. Principal component analysis [PCA] was performed to compare the categorical diet data. Procrustes [ade4 R library] was employed to access correlations between the beta diversity [weighted UniFrac with VST normalisation] and dietary data [PCA]. All p-values were adjusted where necessary using the Benjamini and Hochberg method to correct multiple testing. For linear regression analysis, Spearman correlations were used. Statistical analyses were performed in GraphPad Prism v5, SPSS v15, and R v3.3.0. 3. Results 3.1. Butyryl-CoA:acetate CoA-transferase gene content of microbiota in Crohn’s and ulcerative colitis No major differences for demographic data were found between patients with CD and UC except for previous IBD surgery and 5-aminosalicylic acid compounds usage [Table 1]. Control individuals presented a lower percentage of obese subjects and higher percentage of individuals who had been breastfed and born by caesarean section than patients with IBD. They were also statistically significantly younger than UC patients. Table 1. Clinical profile of study populations. Montreal classification for patients with Crohn’s disease [CD] was collected for 58 subjects. Numbers in parenthesis for medication and Montreal classification correspond to absolute numbers.   CD  UC  Controls  n  71  58  75  Age [years]a  44.3 ± 12.3  48.9 ± 12.0  41.2 ± 16.2  % male  54.9%  65.5%  46.6%  Years since diagnosis  14.2 ± 9.7  16.6 ± 10.1  -  % Previous IBD surgeryb  49.3  0.0  -  Body mass indexc,d  27.7 ± 5.5  26.9 ± 4.3  24.7 ± 3.7  % Obese [BMI > 30]e  27.5  21.4  6.9  % Smokers  8.5  8.5  12.9  % Breastfedf  31.1  32.5  54.0  % Caesarean sectionf  6.5  0.0  14.5  K6 stress scale [K6 ≤ 18]  25.4 ± 4.4 [9%]  26.6 ± 3.1 [4%]  -  Inactive phase  34  27  -  Active phase  16  19  -  Medication  % 5-ASA compoundsb  39.4 [28]  72.4 [42]  -  % Steroids  16.9 [12]  13.8 [8]  -  % Immunosuppressants  25.4 [18]  27.6 [16]  -  % Biologics  28.2 [20]  17.2 [10]  -  Montreal classification  % Age at diagnosis  A1: 6.9 [4]  -  -  A2: 79.3 [46]  A3: 13.8 [8]  % Location  L1: 29.3 [17]  -  -  L2: 25.9 [15]  L3: 43.1 [25]  L1/L4: 1.7 [1]  % Behaviour  B1: 43.1 [25]  -  -  B2: 32.8 [19]  B3: 24.1 [14]  P: 16.7 [9]    CD  UC  Controls  n  71  58  75  Age [years]a  44.3 ± 12.3  48.9 ± 12.0  41.2 ± 16.2  % male  54.9%  65.5%  46.6%  Years since diagnosis  14.2 ± 9.7  16.6 ± 10.1  -  % Previous IBD surgeryb  49.3  0.0  -  Body mass indexc,d  27.7 ± 5.5  26.9 ± 4.3  24.7 ± 3.7  % Obese [BMI > 30]e  27.5  21.4  6.9  % Smokers  8.5  8.5  12.9  % Breastfedf  31.1  32.5  54.0  % Caesarean sectionf  6.5  0.0  14.5  K6 stress scale [K6 ≤ 18]  25.4 ± 4.4 [9%]  26.6 ± 3.1 [4%]  -  Inactive phase  34  27  -  Active phase  16  19  -  Medication  % 5-ASA compoundsb  39.4 [28]  72.4 [42]  -  % Steroids  16.9 [12]  13.8 [8]  -  % Immunosuppressants  25.4 [18]  27.6 [16]  -  % Biologics  28.2 [20]  17.2 [10]  -  Montreal classification  % Age at diagnosis  A1: 6.9 [4]  -  -  A2: 79.3 [46]  A3: 13.8 [8]  % Location  L1: 29.3 [17]  -  -  L2: 25.9 [15]  L3: 43.1 [25]  L1/L4: 1.7 [1]  % Behaviour  B1: 43.1 [25]  -  -  B2: 32.8 [19]  B3: 24.1 [14]  P: 16.7 [9]  IBD, inflammatory bowel disease; UC, ulcerative colitis; BMI, body mass index; 5-ASA, 5-aminosalicylic acid. ap < 0.01 for UC vs controls. bp < 0.001 for CD vs UC. cp < 0.01 for CD vs controls. dp < 0.05 for UC vs controls. ep < 0.01. fp < 0.05. View Large The number of copies of BCoAT gene determined by qPCR did not correlate with age for the whole population or within each group. We observed no correlation with time elapsed after CD/UC diagnosis, either. Although no differences in BCoAT gene content were found between males and females when analysing all samples, there were significant higher BCoAT gene levels for males in UC. Regarding disease, BCoAT gene content was significantly lower in CD samples compared with healthy controls and UC [p < 0.001 and p < 0.05, respectively], whereas no differences were detected between control individuals and UC [Figure 1A; and Supplementary Table 1, available as Supplementary data at ECCO-JCC online]. When the disease activity was considered, BCoAT gene content remained significantly different between controls and both CD groups and also between controls and the UC active group, whereas patients with inactive UC showed higher BCoAT gene levels than both CD groups [Figure 1B; and Supplementary Table 1]. Similar differences between groups were observed when patients with IBD were classified according to faecal calprotectin levels, a specific biomarker of active intestinal inflammation, though no change in BCoAT gene content was observed between healthy controls and patients with UC and high calprotectin concentration [Figure 1C; and Supplementary Table 1]. As expected, concentrations of faecal calprotectin were significantly higher [p < 0.001] for active patients in both CD [mean ± SD: 971 ± 1652 vs 49 ± 58 µg/g] and UC [mean ± SD: 967 ± 1459 vs 57 ± 74 µg/g], demonstrating a good concordance with patient classification following clinical criteria [Figure 1D]. Figure 1. View largeDownload slide Quantification of BCoAT gene content in faecal samples of patients with IBD and healthy controls by qPCR. Tukey boxplot [whiskers represent median ± 1.5 interquartile range] when three groups were considered: CD [n = 71], controls [n = 75], and UC [n = 58] [A]; when patients with IBD were sub-grouped according to disease activity: CD active [n = 16], CD remission [n = 34], controls [n = 75], UC remission [n = 27] and UC active [n = 19] [B]; and when patients with IBD were sub-grouped according to faecal calprotectin [FCAL] concentration: CD and FCAL > 100 [n = 29], CD and FCAL < 100 [n = 42], controls [n = 75], UC and FCAL < 100 [n = 36], and UC and FCAL > 100 [n = 22]. Graph depicting mean calprotectin concentration with whiskers representing standard error of the mean for patients with CD and UC classified according to disease activity [D]. Tukey boxplot for BCoAT gene content in patients with CD classified according to disease location [L1 = ileal, L2 = colonic, L3 = ileo-colonic] [E]; and disease behaviour [B1 = non-stenotic/non-penetrating, B2 = stenotic, B3 = penetrating] [F]. *p < 0.05, **p < 0.01, ***p < 0.001. IBD, inflammatory bowel disease; qPCR, quantitative polymerase chain reaction; CD, Crohn’s disease; UC, ulcerative colitis. Figure 1. View largeDownload slide Quantification of BCoAT gene content in faecal samples of patients with IBD and healthy controls by qPCR. Tukey boxplot [whiskers represent median ± 1.5 interquartile range] when three groups were considered: CD [n = 71], controls [n = 75], and UC [n = 58] [A]; when patients with IBD were sub-grouped according to disease activity: CD active [n = 16], CD remission [n = 34], controls [n = 75], UC remission [n = 27] and UC active [n = 19] [B]; and when patients with IBD were sub-grouped according to faecal calprotectin [FCAL] concentration: CD and FCAL > 100 [n = 29], CD and FCAL < 100 [n = 42], controls [n = 75], UC and FCAL < 100 [n = 36], and UC and FCAL > 100 [n = 22]. Graph depicting mean calprotectin concentration with whiskers representing standard error of the mean for patients with CD and UC classified according to disease activity [D]. Tukey boxplot for BCoAT gene content in patients with CD classified according to disease location [L1 = ileal, L2 = colonic, L3 = ileo-colonic] [E]; and disease behaviour [B1 = non-stenotic/non-penetrating, B2 = stenotic, B3 = penetrating] [F]. *p < 0.05, **p < 0.01, ***p < 0.001. IBD, inflammatory bowel disease; qPCR, quantitative polymerase chain reaction; CD, Crohn’s disease; UC, ulcerative colitis. We also performed statistical analysis for BCoAT gene concentrations among the different categories of Montreal classification in patients with CD. No differences were found for age at diagnosis, but we detected significant changes depending on disease location and behaviour. Patients with ileal location showed significantly lower BCoAT gene levels than those with colonic location [Figure 1E; and Supplementary Table 1]. Regarding behaviour, patients with stenotic phenotype had the more decreased BCoAT gene content, which was significantly lower than BCoAT gene values in patients with non-stenotic/non-penetrating disease [Figure 1F; and Supplementary Table 1]. Otherwise, BCoAT gene concentration was not different between patients with CD who underwent surgery and those with no previous surgery. 3.2. Relationship between patients sub-grouped according to BCoAT gene content and treatment, clinical variables, and inflammation To examine further associations with current treatment and with other variables [disease activity, obesity, previous IBD surgery, and being breastfed], patients with CD and UC classified as active or inactive were sub-grouped each into two categories, using a cut-off of 9.5 [log10] number of copies of BCoAT gene/gram wet faeces. This cut-off value was chosen to achieve a similar number of samples in each category: CD ≤ 9.5 [n = 29], CD > 9.5 [n = 21], UC ≤ 9.5 [n = 19], and UC > 9.5 [n = 27]. In that way, BCoAT level was transformed into a dichotomous nominal variable for statistical analysis [Table 2; and Supplementary Figure 1, available as Supplementary data at ECCO-JCC online]. Univariate analysis showed that low BCoAT levels were associated with active CD, whereas no association between BCoAT gene concentration and any other variable was found in patients with UC. We estimated that patients with active CD had 4.9 times higher probability of presenting BCoAT levels below 9.5 log10 copies/g. In a multivariate logistic regression model considering all four categories of treatment, or all four other variables, disease activity remained as the only variable with statistical significance in patients with CD. Treatment with biologics was associated with low BCoAT gene content in the multivariate analysis for patients with UC, but this observation was affected by a low number of anti-tumour necrosis factor [TNF]-treated UC patients in the analysis and the fact that 62.5% of them were in active stage because of being recruited in an early induction phase. Table 2. Univariate and multivariate logistic regression analysis regarding treatment [A] and other data [disease activity, obesity, previous IBD surgery and being breastfed] [B]. Odds ratio [OR], confidence interval [CI] at 95% and p-value [p] in univariate analysis are referred to present BCoAT levels below 9.5 log10 number of copies of BCoAT gene/g wet faeces; p-value for univariate analysis was calculated using Fisher’s exact test. Surgery was not considered in the analysis of patients with UC. Smoking and caesarean section were not considered as presenting low numbers.   Crohn’s disease  Ulcerative colitis  Univariate  Multivariate  Univariate  Multivariate    OR  CI 95%  p  p  OR  CI 95%  p  p  A. Treatment  5-ASA compounds  0.82  0.3–2.6  0.78  0.73  1.18  0.3–4.4  1.00  0.81  Steroids  5.22  0.6–47.1  0.22  0.11  2.13  0.4–10.9  0.42  0.36  Immunosuppressant  2.24  0.6–8.5  0.34  0.23  1.62  0.4–6.1  0.51  0.48  Biologics  1.44  0.4–5.2  0.75  0.57  5.77  1.0–32.7  0.05  0.03  B. Other data                  Disease active  4.88  1.2–20.3  0.03  0.02  1.53  0.5–5.0  0.55  0.07  BMI > 30  0.28  0.1–1.0  0.06  0.14  0.88  0.2–4.3  1.00  0.73  Previous surgery  1.24  0.4–3.9  0.78  0.86  -  -  -  -  Breastfed  1.13  0.3–4.3  1.00  0.41  0.70  0.2–3.2  0.72  0.64    Crohn’s disease  Ulcerative colitis  Univariate  Multivariate  Univariate  Multivariate    OR  CI 95%  p  p  OR  CI 95%  p  p  A. Treatment  5-ASA compounds  0.82  0.3–2.6  0.78  0.73  1.18  0.3–4.4  1.00  0.81  Steroids  5.22  0.6–47.1  0.22  0.11  2.13  0.4–10.9  0.42  0.36  Immunosuppressant  2.24  0.6–8.5  0.34  0.23  1.62  0.4–6.1  0.51  0.48  Biologics  1.44  0.4–5.2  0.75  0.57  5.77  1.0–32.7  0.05  0.03  B. Other data                  Disease active  4.88  1.2–20.3  0.03  0.02  1.53  0.5–5.0  0.55  0.07  BMI > 30  0.28  0.1–1.0  0.06  0.14  0.88  0.2–4.3  1.00  0.73  Previous surgery  1.24  0.4–3.9  0.78  0.86  -  -  -  -  Breastfed  1.13  0.3–4.3  1.00  0.41  0.70  0.2–3.2  0.72  0.64  IBD, inflammatory bowel disease; 5-ASA, 5-aminosalicylic acid; BMI, body mass index. View Large To further investigate the relationship between BCoAT gene content and inflammation, faecal concentrations of calprotectin were compared in the aforementioned groups. Thereby, when CD patients were sub-grouped according to 9.5 log10 copies BCoAT/g cut-off point, average calprotectin level was higher for patients with CD and low BCoAT gene content [512 ± 1301 vs 112 ± 217 µg/g], although this was not statistically significant [p = 0.08]. We observed that patients with UC displayed no difference in average calprotectin when split by BCoAT gene levels [422 ± 803 vs 440 ± 1178 µg/g]. 3.3. Microbial diversity Comparison of microbiota alpha [intra-individual] diversity across the study populations showed that patients with CD and UC had significantly lower diversity compared with control individuals [Figure 2A and C]. When patients with CD and UC were divided according to BCoAT gene content, a lower diversity [p < 0.05 for chao1 index and 0.05 < p < 0.1 for Shannon index] was observed for CD patients with BCoAT levels below 9.5 log10 copies BCoAT/g [Figure 2B and D]. In contrast, this decrease was not found in patients with UC who displayed a similar level of alpha diversity independently of their BCoAT concentration. A positive correlation was observed between BCoAT gene content and alpha diversity, being strongest for patients with CD and low BCoAT levels [Supplementary Figure 2, available as Supplementary data at ECCO-JCC online]. Figure 2. View largeDownload slide Comparison of microbiota alpha diversity across groups. Tukey plots showing alpha diversity chao1 index for CD, UC and control groups [A] and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [B]. Tukey plots showing alpha diversity Shannon index for CD, UC and control groups [C]; and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [D]. Statistical differences between CD ≤ 9.5 vs UC ≤ 9.5, CD > 9.5 vs UC > 9.5, CD ≤ 9.5 vs UC > 9.5 and CD > 9.5 vs UC ≤ 9.5 are not shown. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Figure 2. View largeDownload slide Comparison of microbiota alpha diversity across groups. Tukey plots showing alpha diversity chao1 index for CD, UC and control groups [A] and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [B]. Tukey plots showing alpha diversity Shannon index for CD, UC and control groups [C]; and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [D]. Statistical differences between CD ≤ 9.5 vs UC ≤ 9.5, CD > 9.5 vs UC > 9.5, CD ≤ 9.5 vs UC > 9.5 and CD > 9.5 vs UC ≤ 9.5 are not shown. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Regarding microbiota beta [inter-individual] diversity, PCoA based on weighted UniFrac distances showed a shift in microbiota composition for patients with CD and UC away that of healthy controls, which was even more pronounced for CD [Figure 3A]. Categorising patients with CD and UC according to BCoAT gene content revealed that patients with CD and low BCoAT gene levels more often displayed a statistically significant shift along principal co-ordinate 1 [PC1] axis compared with healthy controls and UC patients than did those with CD and high BCoAT gene levels, with the latter showing a shift towards control individuals [Figure 3B]. Figure 3. View largeDownload slide Comparison of microbiota beta diversity across groups. Weighted UniFrac VST principal coordinate analysis [PCoA] and violin plots representing PCoA points along PC1 and PC2 axis for CD, UC, and control groups [A]; and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT /g cut-off point [B]. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Figure 3. View largeDownload slide Comparison of microbiota beta diversity across groups. Weighted UniFrac VST principal coordinate analysis [PCoA] and violin plots representing PCoA points along PC1 and PC2 axis for CD, UC, and control groups [A]; and for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT /g cut-off point [B]. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. 3.4. Microbiota compositional changes in known butyrogenic bacteria We investigated whether the observed differences in microbiota composition were due in part to alterations in known butyrate producers, comprising two groups of bacteria [Clostridium clusters IV and XIVa], three genera [Anaerostipes, Butyricicoccus, and Roseburia] and four species [Blautia faecis, Eubacterium hallii, Faecalibacterium prausnitzii, and Ruminococcus torques]. We observed statistically significant changes for Clostridium cluster IV, Roseburia, and Faecalibacterium prausnitzii in the comparison between controls and patients with CD, but no butyrogenic taxa were altered between controls and patients with UC [Supplementary Table 2, available as Supplementary data at ECCO-JCC online]. Those changes were even more evident for patients with CD when considering BCoAT gene content [Supplementary Table 2; and Figure 4]. We found significant reductions for six of the taxa analysed between CD patients with low BCoAT concentration and controls. Furthermore, Butyricicoccus and Roseburia genera [and also Clostridium cluster XIVa and Eubacterium hallii with an adjusted p-value between 0.05 and 0.10] showed a decrease in CD patients with low BCoAT level compared with those with high BCoAT gene content. Accordingly, the only significant difference between CD patients with high BCoAT levels and controls was found for Clostridium cluster IV. In patients with UC, significant changes in bacterial composition between those with low BCoAT gene content and controls were limited to a decrease in Roseburia genus, while no changes were observed when comparing UC patients with high BCoAT levels and controls. Figure 4. View largeDownload slide Tukey boxplots for bacteria composition in healthy control samples and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point. Percentage of each of the group/genus/species considered with respect to total bacteria in stool samples are represented on the Y axis. Groups: Clostridium cluster IV and Clostridium cluster XIVa. Genera: Roseburia, Anaerostipes, and Butyricicoccus. Species: Faecalibacterium prausnitzii and Eubacterium hallii. Statistical differences between CD ≤ 9.5 vs UC ≤ 9.5, CD > 9.5 vs UC > 9.5, CD ≤ 9.5 vs UC > 9.5 and CD > 9.5 vs UC ≤ 9.5 are not shown. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Figure 4. View largeDownload slide Tukey boxplots for bacteria composition in healthy control samples and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point. Percentage of each of the group/genus/species considered with respect to total bacteria in stool samples are represented on the Y axis. Groups: Clostridium cluster IV and Clostridium cluster XIVa. Genera: Roseburia, Anaerostipes, and Butyricicoccus. Species: Faecalibacterium prausnitzii and Eubacterium hallii. Statistical differences between CD ≤ 9.5 vs UC ≤ 9.5, CD > 9.5 vs UC > 9.5, CD ≤ 9.5 vs UC > 9.5 and CD > 9.5 vs UC ≤ 9.5 are not shown. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. 3.5. Influence of diet on butyrate-synthetic capacity Differences in diet were also considered as a factor that could affect butyrate synthesis in the gut by modifying microbial composition [Table 3]. We observed a lower vegetable and fruit intake in patients with CD compared with healthy controls, and they showed increased intake of processed low fibre bread [white bread] and high sugar foods compared with controls. When patients with CD and UC were sub-grouped according to BCoAT gene content, patients with CD and high BCoAT gene concentration had a larger intake of nuts than those with low BCoAT levels, whereas no dietary changes were found for patients with UC. In the comparison of dietary habits among these groups and controls, we observed major significant differences between healthy controls and CD patients with low BCoAT gene content, with the latter showing reduced intake of certain foods containing fibre [Supplementary Table 3, available as Supplementary data at ECCO-JCC online] such as vegetables, fruits, cereals, brown/wholemeal bread, and nuts, and increased intake of high sugar food and white bread. Minor changes were noticed also in the diet of patients with UC and high BCoAT gene content compared with healthy controls. Table 3. Statistically significant changes observed for food categories in diet comparisons from data obtained through a food frequency questionnaire. Mean ± SD [standard deviation] refers to average number of times per month a food in each category was consumed. In the right column, p-values obtained in Wilcoxon test and adjusted for multiple testing are shown. Healthy controls vs CD  Healthy controls Mean ± SD  Patients with CD Mean ± SD  p-Value  Fruit  17.0 ± 17.0  11.0 ± 12.1  < 0.001  Vegetables  36.5 ± 38.3  18.5 ± 13.7  < 0.001  High sugar food  22.5 ± 17.9  29.7 ± 22.0  < 0.001  White bread  4.5 ± 8.1  8.5 ± 10.0  < 0.01  CD > 9.5 vs CD ≤ 9.5 copies BCoAT gene/g  CD > 9.5 copies BCoAT gene/g Mean ± SD  CD ≤ 9.5 copies BCoAT gene/g Mean ± SD  p-value  Nuts  3.9 ± 5.0  0.9 ± 1.9  < 0.01  Healthy controls vs CD ≤ 9.5 copies BCoAT gene/g  Healthy controls Mean ± SD  CD ≤ 9.5 copies BCoAT gene/g Mean ± SD  p-Value  Brown/wholemeal bread  8.5 ± 12.2  4.9 ± 8.2  < 0.05  Fruit  17.0 ± 17.0  7.9 ± 7.6  < 0.001  High fibre cereals  4.6 ± 6.0  2.2 ± 4.0  < 0.01  Nuts  3.7 ± 13.4  0.9 ± 1.9  < 0.05  Vegetables  36.5 ± 38.3  13.9 ± 9.2  < 0.001  High sugar food  22.5 ± 17.9  34.7 ± 20.9  < 0.01  White bread  4.5 ± 8.1  8.2 ± 10.9  < 0.05  Healthy controls vs UC > 9.5 copies BCoAT gene/g  Healthy controls Mean ± SD  UC > 9.5 copies BCoAT gene/g Mean ± SD  p-Value  Fruit  17.0 ± 17.0  11.3 ± 7.5  < 0.05  Vegetables  36.5 ± 38.3  26.6 ± 29.3  < 0.05  High sugar food  22.5 ± 17.9  40.1 ± 30.5  < 0.05  Healthy controls vs CD  Healthy controls Mean ± SD  Patients with CD Mean ± SD  p-Value  Fruit  17.0 ± 17.0  11.0 ± 12.1  < 0.001  Vegetables  36.5 ± 38.3  18.5 ± 13.7  < 0.001  High sugar food  22.5 ± 17.9  29.7 ± 22.0  < 0.001  White bread  4.5 ± 8.1  8.5 ± 10.0  < 0.01  CD > 9.5 vs CD ≤ 9.5 copies BCoAT gene/g  CD > 9.5 copies BCoAT gene/g Mean ± SD  CD ≤ 9.5 copies BCoAT gene/g Mean ± SD  p-value  Nuts  3.9 ± 5.0  0.9 ± 1.9  < 0.01  Healthy controls vs CD ≤ 9.5 copies BCoAT gene/g  Healthy controls Mean ± SD  CD ≤ 9.5 copies BCoAT gene/g Mean ± SD  p-Value  Brown/wholemeal bread  8.5 ± 12.2  4.9 ± 8.2  < 0.05  Fruit  17.0 ± 17.0  7.9 ± 7.6  < 0.001  High fibre cereals  4.6 ± 6.0  2.2 ± 4.0  < 0.01  Nuts  3.7 ± 13.4  0.9 ± 1.9  < 0.05  Vegetables  36.5 ± 38.3  13.9 ± 9.2  < 0.001  High sugar food  22.5 ± 17.9  34.7 ± 20.9  < 0.01  White bread  4.5 ± 8.1  8.2 ± 10.9  < 0.05  Healthy controls vs UC > 9.5 copies BCoAT gene/g  Healthy controls Mean ± SD  UC > 9.5 copies BCoAT gene/g Mean ± SD  p-Value  Fruit  17.0 ± 17.0  11.3 ± 7.5  < 0.05  Vegetables  36.5 ± 38.3  26.6 ± 29.3  < 0.05  High sugar food  22.5 ± 17.9  40.1 ± 30.5  < 0.05  CD, Crohn’s disease; UC, ulcerative colitis. View Large In the PCA analysis, we detected that the diet of patients with CD exhibited a pronounced shift away from the healthy controls’ diet, whereas the diet of patients with UC displayed a lesser shift but still statistically significant for PC2 axis [Figure 5A]. Interestingly, the diet of patients with CD and higher BCoAT gene content, but not for UC, exhibited a shift towards healthy controls’ diet in the PCA plot and showed lower statistical difference compared with controls [p < 0.05] than did patients with CD and low BCoAT gene levels [p < 0.001] in the PC2 axis [Figure 5B]. No major differences were observed either in the statistical analysis or in the PCA plot for the comparison between the 20 healthy controls with the highest BCoAT content and the 20 controls with the lowest BCoAT level [Figure 5C]. We investigated also if patients with CD showed differences in diet according to their disease location or behaviour, but we found no significant differences in the statistical analysis and no major shifts in the PCA plot [Supplementary Figure 3, available as Supplementary data at ECCO-JCC online]. Figure 5. View largeDownload slide Comparison of dietary data by principal component analysis [PCA] and by Wilcoxon tests [for PCA points along PC1 and PC2 axis, which were depicted as violin plots] for CD [n = 63], UC [n = 56], and control [n = 67] groups [A]; for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [CD ≤ 9.5 n = 27, CD > 9.5 n = 17, UC ≤ 9.5 n = 18, UC > 9.5 n = 27] [B]; and for the 20 healthy controls with the highest BCoAT gene content and the 20 healthy controls with the lowest BCoAT concentration [C]. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Figure 5. View largeDownload slide Comparison of dietary data by principal component analysis [PCA] and by Wilcoxon tests [for PCA points along PC1 and PC2 axis, which were depicted as violin plots] for CD [n = 63], UC [n = 56], and control [n = 67] groups [A]; for control group and patients with CD/UC divided according to 9.5 log10 copies BCoAT/g cut-off point [CD ≤ 9.5 n = 27, CD > 9.5 n = 17, UC ≤ 9.5 n = 18, UC > 9.5 n = 27] [B]; and for the 20 healthy controls with the highest BCoAT gene content and the 20 healthy controls with the lowest BCoAT concentration [C]. *p < 0.05, **p < 0.01, ***p < 0.001. CD, Crohn’s disease; UC, ulcerative colitis. Finally, procrustes analysis of the relationship between microbiota composition and diet revealed a significant correlation [p < 0.001], suggesting that diet had a direct or indirect effect on microbiota composition [Supplementary Figure 4, available as Supplementary data at ECCO-JCC online]. 4. Discussion Our results showed a reduction in the genetic capacity of colonic microbiota to produce butyrate, which is not only more pronounced in CD than in UC, but it also appears more relevant for CD activity, as low BCoAT gene content was found significantly associated only with active CD. Within CD patients, ileal location and stenotic behaviour showed the lowest BCoAT gene concentrations, being significantly decreased compared with colonic location and non-stenotic/non-penetrating behaviour, respectively. In accordance with this, the inflammation levels, disease-related changes in microbiota composition, and decreased percentage of butyrate-producers were greater in patients with CD having low BCoAT gene content. In addition, we noted dietary differences between healthy individuals and CD patients with low BCoAT gene content, indicating a decreased intake of fibre-rich food [like vegetables, fruits, or cereals] for the latter. Taken together, these results confirm existing knowledge about a decrease in butyrate-producer species in IBD, and contribute new evidence supporting the idea that a reduction in butyrate synthesis genetic capacity is more pronounced and relevant for CD than for UC, with diet being a plausible determinant behind butyrate decline. We observed a significant reduction in BCoAT gene levels for active and inactive patients with CD and active patients with UC as compared with controls. It constitutes a novel approach for evaluating butyrate alterations in IBD, since a dozen of previous studies were focused on faecal butyrate determination by metabolomic approaches. Although most of the studies showed a decrease in butyrate levels in patients with IBD, their results were discrepant. Some reported greater reductions for patients with CD than in UC,13 or only for active Crohn’s.11,14,33 Others described differences between patients with UC and healthy controls,14,34 but some reports were conflicting and found changes only for patients with active UC35 or did not observe any such differences.36 Although the concentration of butyrate in faecal samples may reflect the relative abundance of butyrate-producing bacteria in the gut, chromatographic measurements have limitations.37 First, they may be heavily influenced by butyrate uptake by the host. Second, physiological processes in the gut such as binding, degradation, and mucosal absorption, hamper the reliable estimation of butyrate concentration by direct analysis of faecal specimens. Furthermore, the diverse techniques employed [mainly HPLC, GC or GC/MS] coupled with different detector types and different extraction and derivatisation procedures confound standardisation. In contrast, qPCR is more standardised than the diverse chromatographic techniques. Therefore, we believe that this qPCR analysis of BCoAT gene content represents a reliable and useful method for evaluating butyrate synthesis capacity of gut microbes, which could complement chromatographic determinations. For example, a recent study showed significant differences in BCoAT gene content associated with a microbial shift between patients with UC responding or not responding to faecal microbiota transplantation, whereas no changes were observed in faecal butyrate levels by HPLC.38 Some differences in baseline characteristics were found between healthy control volunteers and our IBD cohort for age, obesity, breastfeeding, and caesarean birth. We have analysed these variables to check whether they could have an effect on BCoAT gene content. Age was not correlated with BCoAT levels and obesity and breastfeeding showed no association with BCoAT subgroups in both univariate and multivariate logistic regression analysis. In addition, the higher percentage of overweight individuals in our IBD cohort compared with healthy controls seems to be in agreement with recent epidemiological data.39 Birth by caesarean section was not considered, due to low numbers in each group for comparison, but it was described that it only affects microbiota during the first 6 months of life.40 Our data indicated a significant decrease in microbiota alpha diversity and a relevant shift in beta diversity away from healthy controls for patients with CD and UC, in agreement with previous reports.15 Both changes in diversity were more prominent for patients with CD and, only in them we observed a higher microbiota alpha diversity and a more similar microbiota to that of controls in patients with high BCoAT gene concentration. Our analysis of changes in particular butyrogenic taxa also proved that such alterations were smaller in patients with CD and high BCoAT gene content. Within the Firmicutes phylum, two 16S-rRNA-defined phylogenetics groups, Clostridium cluster IV and XIVa, include the majority of the bacteria identified as capable of synthesising butyrate. A previous study of butyrogenic bacteria diversity in human faecal samples using BCoAT amplification revealed the importance of the Roseburia group and E. hallii, with an unexpected under-representation for F. prausnitzii which could be explained by a selective loss of F. prausnitzii BCoAT sequences during the cloning step.41 Strain SS2-1 that was reported in this work as a major butyrate-producer using the BCoAT enzyme, was later identified as Anaerostipes hadrus.42 All these taxa were selected for comparison [relative abundance] in the stool microbiota, jointly with genus Butyricicoccus20 and B. faecis and R. torques species.43 We observed reduced relative abundances for six taxa in patients with CD and low BCoAT content compared with healthy controls, whereas solely one taxon was decreased for the same BCoAT sub-group of patients with UC. As only one taxon, Clostridium cluster IV, was declined for patients with CD and high BCoAT gene concentration when compared with controls, increased BCoAT levels correlated with a relevant improvement in butyrogenic bacteria abundance. Our findings for butyrate-producers in patients with CD are in general agreement with previous reports on this topic.44 Several studies, primarily in UC, described promising but inconclusive results in ameliorating inflammation and symptoms with therapeutic administration of butyrate.10 As our results showed no difference in BCoAT gene content between patients with inactive UC and healthy controls, it could be hypothesised that interventions aimed to increase butyrate levels in their colon would result in minimal effects. A previous interventional study confirmed this hypothesis, since administration of rectal butyrate enemas for 3 weeks in patients with inactive UC caused minor improvements on inflammatory and oxidative stress measurements.45 Butyrate was reported as possibly effective in achieving CD remission when administered orally.46,47 The alterations in the microbiota and the reduction of known butyrate producers in patients with CD, which were higher in those with low BCoAT gene content, suggest that microbiota modulation, either directly or promoted through prebiotics, may provide more consistent and efficient changes in butyrate concentrations in the gut compared with direct butyrate administration. Prebiotic compounds described as stimulators of butyrate synthesis are resistant starch, oat bran, sorbitol, galacturonic and glucuronic acid, inulin, and fructo-oligosaccharides.48 Regarding probiotics, promising results for butyrate producers that could be used as pharmabiotics by themselves to attenuate IBD were obtained for F. prausnitzii, and its culture supernatant,49–52 and B. paellicorum,20,53 although negative effects were also reported for the inoculation of the butyrate producer strain Anaerostipes hadrus BPB5 in a DSS-induced colitis mouse model.54 Another means to achieve more long-term changes in the gut microbiota and thus in butyrate levels might be diet. The impact of diet on butyrate and other SCFA produced by the microbiota has been widely investigated, in both epidemiological and interventional studies.55 For example, European children had almost four times lower levels of butyrate than children living in a rural Burkina Faso village, suggesting that a diet rich in fibre and low in sugar and fat could favour SCFA-producing bacteria.56 Therefore, it is not surprising that dietary factors related to SCFA could affect the course and treatment of IBD.57 Our procrustes analysis showed that diet had a noticeable effect on the microbiota. In particular, patients with CD and low BCoAT gene content had a more ‘westernised’ diet with lower intake of fruit, high fibre cereals, nuts, brown/wholemeal bread, and vegetables compared with controls, and also a higher intake of high sugar foods and white bread, which are sources of carbohydrates. Moreover, the food group which includes nuts, which tended to be consumed to a greater degree by CD patients with higher BCoAT levels than by those with lower levels of BCoAT, may represent an alternative source of fibre to increase butyrate. Of course, recommendations regarding dietary fibre must be judicious particularly in patients at risk of stricture formation, where reduced dietary fibre is a traditional recommendation. There were some limitations in our study. First, although disease activity was not confirmed by endoscopy, a notional gold standard, we supplemented the physicians’ assessment with conventional clinical indices and faecal calprotectin. Second, the primers employed were designed based on a limited number of sequences, but they were proved to capture the known butyrate producers with higher compositional representation in the gut microbiota.41 Third, our analysis of butyrate-producing capacity was limited to BCoAT gene content, although other factors are known to affect butyrate availability and utilisation; for example, butyrate uptake/transportation by colonocytes may be impaired in IBD.58 In addition, butyrate synthesis through the butyrate kinase pathway or from other alternative sources, such as lysine or 4-aminobutyrate/succinate, were not considered, but these pathways appear to be much less important from a quantitative point of view.59 Regarding dietary data, the FFQ employed did not provide information of portion sizes, which did not allow calculation of dietary intake for nutrients. It is beyond the aim of the present study evaluate whether reduced butyrate synthesis capacity is cause or consequence of the disease, even though we hypothesise that normalisation of the gut microbiota butyrogenic capacity may associate with an improvement in patients with CD. New studies will ultimately be required to investigate whether BCoAT gene content determination by qPCR could be helpful for patient management in IBD. In conclusion, our study confirms evidence for butyrate alterations in IBD and extends current knowledge by showing that a decrease in the genetic capacity of the gut microbiota to synthesise butyrate is associated with active disease, higher degree of inflammation, and greater changes in the microbiota in patients with CD but not in those with UC. Since dietary habits influence the butyrate synthetic capacity of the microbiota, the judicious use of dietary fibre deserves reconsideration as a therapeutic strategy. Funding This work was supported by Science Foundation Ireland [SFI] under Grant Numbers 11/SIRG/B2162 and SFI/12/RC/2273. EJLM conducted part of this research under a contract funded mostly with a research grant from the European Crohn’s and Colitis Organization [year 2014]. Conflict of Interest APC Microbiome Institute receives funding from several private companies, but none of them has been involved in the present study. Author Contributions EJLM, AGC, MJC: study concept and data analysis; EJLM, AGC, FS, MJC: output formatting and manuscript writing; EJLM: sample processing and laboratory work; EJLM, JFCG: performing calprotectin assay; EJLM, AGC, CM, DS, CGMG, SAJ, FS, MJC: patient recruitment and data collection; JN, CH: recruitment of healthy controls; all authors: final review of manuscript. All authors have approved the final version of the manuscript. Supplementary Data Supplementary data are available at ECCO-JCC online. Acknowledgments We would like to thank Dr Feargal Ryan, Ms Emily Power, Mr Alvaro Lopez-Gallardo, and Dr Kalaimathi Govindarajan [University College Cork] for helping in metadata processing, Ms Ángeles Cabezas-Martínez and Mrs María Jesús Rocha-Bogas [Complejo Hospitalario de Toledo] for assisting in calprotectin assay, Dr Fabien Cousin and Dr Guillaume Borrel [former post-doctoral researchers at University College Cork] for helping in setting up qPCR assay, Dr Janette Walton [School of Food and Nutritional Science, University College Cork] for collaborating in dietary data interpretation, and Ms Margot Hurley and Ms Catherine O’Riordan [Cork University Hospital] for collecting stool samples and providing questionnaires. References 1. Sheehan D, Moran C, Shanahan F. The microbiota in inflammatory bowel disease. J Gastroenterol  2015; 50: 495– 507. Google Scholar CrossRef Search ADS PubMed  2. Louis P, Duncan SH, McCrae SI, Millar J, Jackson MS, Flint HJ. Restricted distribution of the butyrate kinase pathway among butyrate-producing bacteria from the human colon. J Bacteriol  2004; 186: 2099– 106. Google Scholar CrossRef Search ADS PubMed  3. Hamer HM, Jonkers D, Venema K, Vanhoutvin S, Troost FJ, Brummer RJ. Review article: the role of butyrate on colonic function. Aliment Pharmacol Ther  2008; 27: 104– 19. Google Scholar CrossRef Search ADS PubMed  4. Plöger S, Stumpff F, Penner GBet al.   Microbial butyrate and its role for barrier function in the gastrointestinal tract. Ann N Y Acad Sci  2012; 1258: 52– 9. Google Scholar CrossRef Search ADS PubMed  5. Zhang LS, Davies SS. Microbial metabolism of dietary components to bioactive metabolites: opportunities for new therapeutic interventions. Genome Med  2016; 8: 46. Google Scholar CrossRef Search ADS PubMed  6. Segain JP, Raingeard de la Blétière D, Bourreille Aet al.   Butyrate inhibits inflammatory responses through NFkappaB inhibition: implications for Crohn’s disease. Gut  2000; 47: 397– 403. Google Scholar CrossRef Search ADS PubMed  7. Arpaia N, Campbell C, Fan Xet al.   Metabolites produced by commensal bacteria promote peripheral regulatory T-cell generation. Nature  2013; 504: 451– 5. Google Scholar CrossRef Search ADS PubMed  8. Furusawa Y, Obata Y, Fukuda Set al.   Commensal microbe-derived butyrate induces the differentiation of colonic regulatory T cells. Nature  2013; 504: 446– 50. Google Scholar CrossRef Search ADS PubMed  9. Smith PM, Howitt MR, Panikov Net al.   The microbial metabolites, short-chain fatty acids, regulate colonic Treg cell homeostasis. Science  2013; 341: 569– 73. Google Scholar CrossRef Search ADS PubMed  10. Sossai P. Butyric acid: what is the future for this old substance? Swiss Med Wkly  2012; 142: w13596. Google Scholar PubMed  11. De Preter V, Joossens M, Ballet Vet al.   Metabolic profiling of the impact of oligofructose-enriched inulin in Crohn’s disease patients: a double-blinded randomized controlled trial. Clin Transl Gastroenterol  2013; 4: e30. Google Scholar CrossRef Search ADS PubMed  12. De Preter V, Machiels K, Joossens Met al.   Faecal metabolite profiling identifies medium-chain fatty acids as discriminating compounds in IBD. Gut  2015; 64: 447– 58. Google Scholar CrossRef Search ADS PubMed  13. Marchesi JR, Holmes E, Khan Fet al.   Rapid and noninvasive metabonomic characterization of inflammatory bowel disease. J Proteome Res  2007; 6: 546– 51. Google Scholar CrossRef Search ADS PubMed  14. Takaishi H, Matsuki T, Nakazawa Aet al.   Imbalance in intestinal microflora constitution could be involved in the pathogenesis of inflammatory bowel disease. Int J Med Microbiol  2008; 298: 463– 72. Google Scholar CrossRef Search ADS PubMed  15. Kostic AD, Xavier RJ, Gevers D. The microbiome in inflammatory bowel disease: current status and the future ahead. Gastroenterology  2014; 146: 1489– 99. Google Scholar CrossRef Search ADS PubMed  16. Frank DN, St Amand AL, Feldman RA, Boedeker EC, Harpaz N, Pace NR. Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases. Proc Natl Acad Sci U S A  2007; 104: 13780– 5. Google Scholar CrossRef Search ADS PubMed  17. Cao Y, Shen J, Ran ZH. Association between Faecalibacterium prausnitzii reduction and inflammatory bowel disease: a meta-analysis and systematic review of the literature. Gastroenterol Res Pract  2014; 2014: 872725. Google Scholar PubMed  18. Rajilić-Stojanović M, Shanahan F, Guarner F, de Vos WM. Phylogenetic analysis of dysbiosis in ulcerative colitis during remission. Inflamm Bowel Dis  2013; 19: 481– 8. Google Scholar CrossRef Search ADS PubMed  19. Gevers D, Kugathasan S, Denson LAet al.   The treatment-naive microbiome in new-onset Crohn’s disease. Cell Host Microbe  2014; 15: 382– 92. Google Scholar CrossRef Search ADS PubMed  20. Eeckhaut V, Machiels K, Perrier Cet al.   Butyricicoccus pullicaecorum in inflammatory bowel disease. Gut  2013; 62: 1745– 52. Google Scholar CrossRef Search ADS PubMed  21. Morgan XC, Tickle TL, Sokol Het al.   Dysfunction of the intestinal microbiome in inflammatory bowel disease and treatment. Genome Biol  2012; 13: R79. Google Scholar CrossRef Search ADS PubMed  22. Erickson AR, Cantarel BL, Lamendella Ret al.   Integrated metagenomics/metaproteomics reveals human host-microbiota signatures of Crohn’s disease. PLoS One  2012; 7: e49138. Google Scholar CrossRef Search ADS PubMed  23. Claesson MJ, Jeffery IB, Conde Set al.   Gut microbiota composition correlates with diet and health in the elderly. Nature  2012; 488: 178– 84. Google Scholar CrossRef Search ADS PubMed  24. Harvey RF, Bradshaw JM. A simple index of Crohn’s-disease activity. Lancet  1980; 1: 514. Google Scholar CrossRef Search ADS PubMed  25. Powell-Tuck J, Day DW, Buckell NA, Wadsworth J, Lennard-Jones JE. Correlations between defined sigmoidoscopic appearances and other measures of disease activity in ulcerative colitis. Dig Dis Sci  1982; 27: 533– 7. Google Scholar CrossRef Search ADS PubMed  26. Walmsley RS, Ayres RC, Pounder RE, Allan RN. A simple clinical colitis activity index. Gut  1998; 43: 29– 32. Google Scholar CrossRef Search ADS PubMed  27. Clooney AG, Bernstein CN, Leslie WDet al.   A comparison of the gut microbiome between long-term users and non-users of proton pump inhibitors. Aliment Pharmacol Ther  2016; 43: 974– 84. Google Scholar CrossRef Search ADS PubMed  28. Klindworth A, Pruesse E, Schweer Tet al.   Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res  2013; 41: e1. Google Scholar CrossRef Search ADS PubMed  29. Louis P, Flint HJ. Development of a semiquantitative degenerate real-time pcr-based assay for estimation of numbers of butyryl-coenzyme A [CoA] CoA transferase genes in complex bacterial samples. Appl Environ Microbiol  2007; 73: 2009– 12. Google Scholar CrossRef Search ADS PubMed  30. Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics  2011; 27: 2194– 200. Google Scholar CrossRef Search ADS PubMed  31. Schloss PD, Westcott SL, Ryabin Tet al.   Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol  2009; 75: 7537– 41. Google Scholar CrossRef Search ADS PubMed  32. Allard G, Ryan FJ, Jeffery IB, Claesson MJ. SPINGO: a rapid species-classifier for microbial amplicon sequences. BMC Bioinformatics  2015; 16: 324. Google Scholar CrossRef Search ADS PubMed  33. Bjerrum JT, Wang Y, Hao Fet al.   Metabonomics of human fecal extracts characterize ulcerative colitis, Crohn’s disease and healthy individuals. Metabolomics  2015; 11: 122– 33. Google Scholar CrossRef Search ADS PubMed  34. Huda-Faujan N, Abdulamir AS, Fatimah ABet al.   The impact of the level of the intestinal short chain fatty acids in inflammatory bowel disease patients versus healthy subjects. Open Biochem J  2010; 4: 53– 8. Google Scholar CrossRef Search ADS PubMed  35. Kumari R, Ahuja V, Paul J. Fluctuations in butyrate-producing bacteria in ulcerative colitis patients of North India. World J Gastroenterol  2013; 19: 3404– 14. Google Scholar CrossRef Search ADS PubMed  36. Machiels K, Joossens M, Sabino Jet al.   A decrease of the butyrate-producing species Roseburia hominis and Faecalibacterium prausnitzii defines dysbiosis in patients with ulcerative colitis. Gut  2014; 63: 1275– 83. Google Scholar CrossRef Search ADS PubMed  37. Verbeke KA, Boobis AR, Chiodini Aet al.   Towards microbial fermentation metabolites as markers for health benefits of prebiotics. Nutr Res Rev  2015; 28: 42– 66. Google Scholar CrossRef Search ADS PubMed  38. Fuentes S, Rossen NG, van der Spek MJet al.   Microbial shifts and signatures of long-term remission in ulcerative colitis after faecal microbiota transplantation. ISME J  2017; 11: 1877– 89. Google Scholar CrossRef Search ADS PubMed  39. Moran C, Sheehan D, Shanahan F. The changing phenotype of inflammatory bowel disease. Gastroenterol Res Pract  2016; 2016: 1619053. Google Scholar CrossRef Search ADS PubMed  40. Rutayisire E, Huang K, Liu Y, Tao F. The mode of delivery affects the diversity and colonization pattern of the gut microbiota during the first year of infants’ life: a systematic review. BMC Gastroenterol  2016; 16: 86. Google Scholar CrossRef Search ADS PubMed  41. Louis P, Young P, Holtrop G, Flint HJ. Diversity of human colonic butyrate-producing bacteria revealed by analysis of the butyryl-CoA:acetate CoA-transferase gene. Environ Microbiol  2010; 12: 304– 14. Google Scholar CrossRef Search ADS PubMed  42. Allen-Vercoe E, Daigneault M, White Aet al.   Anaerostipes hadrus comb. nov., a dominant species within the human colonic microbiota; reclassification of Eubacterium hadrum Moore et al. 1976. Anaerobe  2012; 18: 523– 9. Google Scholar CrossRef Search ADS PubMed  43. Takahashi K, Nishida A, Fujimoto Tet al.   Reduced abundance of butyrate-producing bacteria species in the fecal microbial community in Crohn’s Disease. Digestion  2016; 93: 59– 65. Google Scholar CrossRef Search ADS PubMed  44. Wright EK, Kamm MA, Teo SM, Inouye M, Wagner J, Kirkwood CD. Recent advances in characterizing the gastrointestinal microbiome in Crohn’s disease: a systematic review. Inflamm Bowel Dis  2015; 21: 1219– 28. Google Scholar CrossRef Search ADS PubMed  45. Hamer HM, Jonkers DM, Vanhoutvin SAet al.   Effect of butyrate enemas on inflammation and antioxidant status in the colonic mucosa of patients with ulcerative colitis in remission. Clin Nutr  2010; 29: 738– 44. Google Scholar CrossRef Search ADS PubMed  46. Di Sabatino A, Morera R, Ciccocioppo Ret al.   Oral butyrate for mildly to moderately active Crohn’s disease. Aliment Pharmacol Ther  2005; 22: 789– 94. Google Scholar CrossRef Search ADS PubMed  47. Russo I, Luciani A, De Cicco P, Troncone E, Ciacci C. Butyrate attenuates lipopolysaccharide-induced inflammation in intestinal cells and Crohn’s mucosa through modulation of antioxidant defense machinery. PLoS One  2012; 7: e32841. Google Scholar CrossRef Search ADS PubMed  48. Li RW, Li CJ. Enhancing butyrate biosynthesis in the gut for health benefits. In: Butyrate: Food Sources, Functions and Health Benefits. Hauppauge , NY: USA Nova Science Publishers; 2014. 49. Miquel S, Martín R, Rossi Oet al.   Faecalibacterium prausnitzii and human intestinal health. Curr Opin Microbiol  2013; 16: 255– 61. Google Scholar CrossRef Search ADS PubMed  50. Qiu X, Zhang M, Yang X, Hong N, Yu C. Faecalibacterium prausnitzii upregulates regulatory T cells and anti-inflammatory cytokines in treating TNBS-induced colitis. J Crohns Colitis  2013; 7: e558– 68. Google Scholar CrossRef Search ADS PubMed  51. Martín R, Chain F, Miquel Set al.   The commensal bacterium Faecalibacterium prausnitzii is protective in DNBS-induced chronic moderate and severe colitis models. Inflamm Bowel Dis  2014; 20: 417– 30. Google Scholar CrossRef Search ADS PubMed  52. Sokol H, Pigneur B, Watterlot Let al.   Faecalibacterium prausnitzii is an anti-inflammatory commensal bacterium identified by gut microbiota analysis of Crohn disease patients. Proc Natl Acad Sci U S A  2008; 105: 16731– 6. Google Scholar CrossRef Search ADS PubMed  53. Eeckhaut V, Ducatelle R, Sas B, Vermeire S, Van Immerseel F. Progress towards butyrate-producing pharmabiotics: Butyricicoccus pullicaecorum capsule and efficacy in TNBS models in comparison with therapeutics. Gut  2014; 63: 367. Google Scholar CrossRef Search ADS PubMed  54. Zhang Q, Wu Y, Wang Jet al.   Accelerated dysbiosis of gut microbiota during aggravation of DSS-induced colitis by a butyrate-producing bacterium. Sci Rep  2016; 6: 27572. Google Scholar CrossRef Search ADS PubMed  55. Ríos-Covián D, Ruas-Madiedo P, Margolles A, Gueimonde M, de Los Reyes-Gavilán CG, Salazar N. Intestinal short chain fatty acids and their link with diet and human health. Front Microbiol  2016; 7: 185. Google Scholar CrossRef Search ADS PubMed  56. De Filippo C, Cavalieri D, Di Paola Met al.   Impact of diet in shaping gut microbiota revealed by a comparative study in children from Europe and rural Africa. Proc Natl Acad Sci U S A  2010; 107: 14691– 6. Google Scholar CrossRef Search ADS PubMed  57. Lee D, Albenberg L, Compher Cet al.   Diet in the pathogenesis and treatment of inflammatory bowel diseases. Gastroenterology  2015; 148: 1087– 106. Google Scholar CrossRef Search ADS PubMed  58. Thibault R, Blachier F, Darcy-Vrillon B, de Coppet P, Bourreille A, Segain JP. Butyrate utilization by the colonic mucosa in inflammatory bowel diseases: a transport deficiency. Inflamm Bowel Dis  2010; 16: 684– 95. Google Scholar CrossRef Search ADS PubMed  59. Vital M, Howe AC, Tiedje JM. Revealing the bacterial butyrate synthesis pathways by analyzing [meta]genomic data. MBio  2014; 5: e00889. Google Scholar CrossRef Search ADS PubMed  Copyright © 2017 European Crohn’s and Colitis Organisation (ECCO). Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com

Journal

Journal of Crohn's and ColitisOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off