Characteristics of mucosa-associated gut microbiota during treatment in Crohn's disease

Cong He, Huan Wang, Wang-Di Liao, Chao Peng, Xu Shu, Xuan Zhu, Zhen-Hua Zhu

Abstrac t BACKGROUND The dysbiosis of the gut microbiome is evident in Crohn's disease (CD) compared with healthy controls (HC), although the alterations from active CD to remission after treatment are unclear.AIM To characterize the mucosa-associated gut microbiota in patients with CD before and after the induction therapy.METHODS The basic information was collected from the subjects and the CD activity index(CDAI) was calculated in patients. A 16S r RNA sequencing approach was applied to determine the structures of microbial communities in mucosal samples including the terminal ileal, ascending colon, descending colon and rectum. The composition and function of mucosa-associated gut microbiota were compared between samples from the same cohort of patients before and after treatment.Differential taxa were identified to calculate the microbial dysbiosis index (MDI)and the correlation between MDI and CDAI was analyzed using Pearson correlation test. Predictive functional profiling of microbial communities was obtained with PICRUSt.RESULTS There were no significant differences in microbial richness among the four anatomical sites in individuals. Compared to active disease, the alpha diversity of CD in remission was increased towards the level of HC compared to the active stage. The principal coordinate analysis revealed that samples of active CD were clearly separated from those in remission, which clustered close to HC. Sixty-five genera were identified as differentially abundant between active and quiescent CD, with a loss of Fusobacterium and a gain of potential beneficial bacteria including Lactobacillus, Akkermansia, Roseburia, Ruminococcus and Lachnospira after the induction of remission. The combination of these taxa into a MDI showed a positive correlation with clinical disease severity and a negative correlation with species richness. The increased capacity for the inferred pathways including Lipopolysaccharide biosynthesis and Lipopolysaccharide biosynthesis proteins in patients before treatment negatively correlated with the abundance of Roseburia,Ruminococcus and Lachnospira.CONCLUSION The dysbiosis of mucosa-associated microbiota was associated with the disease phenotype and may become a potential diagnostic tool for the recurrence of disease.

Key words: Crohn's disease; Mucosa-associated gut microbiota; Active; Remission; 16S rRNA sequence conflict of interest exists.

INTRODUCTION

Crohn's disease (CD), a subtype of inflammatory bowel disease (IBD), has become a global d isease w ith accelerating incid ence over the last few d ecad es[1]. CD is characterized by multip le episod es of exacerbation and remission, w ith clinical manifestations of d iarrhea, abdominal pain, fistulas and perianal lesions, w hich may affect the whole digestive tract and cause systemic symptoms[2]. Current therapies for IBD includ e anti-inflammatory and immunomod ulatory treatments such as 5-aminosalicylates, corticosteroid s, thiopurines, thalidomide and anti-tumor necrosis factor alpha, all of which aim to achieve clinical remission and mucosal healing[3,4].The pathogenesis of CD is multifactorial and involves the interplay of host genetics,immune d ysregulation and environmental factors resulting in an aberrant immune response and subsequent intestinal inflammation[5].

Recent p rogress in und erstand ing the comp osition and function of human microbiota has revealed the important role of microbiota in immune homeostasis[6].Accumulating stud ies using culture-ind ep end ent techniques have show n the d ysbiosis of gut microbiota in p atients w ith CD, includ ing d ecreased bacterial diversity, with an expansion of putative aggressive groups (such as Enterobacteriaceae,Fusobacterium) combined with decreases in protective groups (such as Faecalibacterium,Roseburia)[7,8]. In add ition to observing the characteristics of gut microbiota in CD, a study by Wang et al[9]evaluated their dynamic changes after infliximab (IFX) therapy and found that the d ysbiosis could be corrected in p atients w ith a sustained therap eutic resp onse. Furthermore, a p rosp ectiv e stud y assessed the stool metagenomes of IBD patients starting biologic therapy and demonstrated a higher abund ance of butyrate prod ucers at baseline in therap y-resp onsive CD p atients,indicating the predicative effect of the gut microbiome in treatment response[10]. Due to the d ysregulated microbiota in the p athogenesis of IBD, several stud ies have reported the potential effect of restoring dysbiotic gut microbiota, including the use of probiotics and unprocessed donor feces, in the management of IBD[11]. It is necessary to clarify the key bacteria that play a role in disease remission and relapse, and then,precise manipulation of these bacteria may become a therapeutic target in the future.

To date, most stud ies investigating the gut microbiota of CD have typically used fecal samples since they are readily obtained[7-9]. How ever, the comp osition of fecal microbiota has been shown to be significantly different from mucosal microbiota; this difference is believed to directly affect epithelial and mucosal function and to be more d eeply involved in the p athop hysiology of CD[12,13]. To collect sufficient mucosal samples, we processed endoscopically uninflamed mucosa, which w as thicker than the inflamed mucosa and p robably more ap propriate for microbial analysis[14]. In addition, only limited differences in microbiota composition w ere observed betw een inflamed and uninflamed mucosa in patients[15]. The previous cross-sectional study of the alterations betw een CD p atients and healthy controls (HC) could be misread based on the interind ivid ual differences, w hich make it d ifficult to characterize the critical bacteria in CD. Thus, we investigated the mucosal-associated microbiome in paired samples from CD patients before and after clinical treatment by 16S r RNA gene sequencing to d etermine the association betw een gut microbiota and d isease activity.

MATERIALS AND METHODS

Study cohort

A prospective study w as performed in nine CD patients who were enrolled in flare at baseline and then ind uced remission after clinical therapy. Inclusion criteria w ere a diagnosis of CD confirmed by endoscopy and histology and the activity of the disease was measured by the CD activity index (CDAI). Six HC w ithout previous history of chronic d isease w ere also recruited in the study from the First Affiliated Hospital of Nanchang University, China. Εxclusion criteria for the tw o groups included severe concomitant disease involving the liver, heart, lung or kidney, pregnancy or breastfeeding, and treatment with antibiotics and prebiotics during the previous 4 w k. The mucosal samples w ere collected d uring the colonoscopy and both the patients and healthy subjects und erw ent intestinal w ashing before the examination. We d id not collect both inflamed and noninflamed tissues since a previous study showed that the mucosal microbiota of inflamed and noninflamed regions of the gastrointestinal tract in CD or ulcerative colitis (UC) w ere ind istinguishable, w ith virtually no taxa d emonstrating d isp rop ortional abund ances at a significant threshold nor any significant diversity differences observed[15]. Written informed consent w as obtained from all the subjects and this study was approved by the Medical Εthics Committee of Nanfang Hospital.

Sample collection and DNA extraction

A total of 74 mucosal biopsies were collected from 15 participants, including 9 patients with CD and 6 healthy individuals. Specimens of terminal ileum, ascending colon, descending colon and rectum in noninflamed mucosa were taken during colonoscopic examination. Sampling included both active and remission stages for each patient who underwent clinical treatment. All the samples were immediately put in liquid nitrogen and stored at - 80 °C before processing.

Microbial DNA was extracted from the mucosal biopsies using the Ε.Z.N.A. stool DNA kit (Omega Biotek, Norcross, GA, United States) accord ing to the manufacturer's protocols. The 16S rDNA V3-V4 region of the Εukaryotic ribosomal RNA gene was amplified by PCR (95 °C for 2 min, followed by 27 cycles at 98 °C for 10 s, 62 °C for 30 s, and 68 °C for 30 s and a final extension at 68 °C for 10 min) using primers 341F: CCTACGGGNGGCWGCAG; 806R: GGACTACHVGGGTATCTAAT,where the barcode is an eight-base sequence unique to each sample. PCRs were performed in triplicate 50 μL mixture containing 5 μL of 10 × KOD Buffer, 5 μL of 2.5 mM d NTPs, 1.5 μL of each primer (5 μM), 1 μL of KOD Polymerase, and 100 ng of template DNA.

16S rRNA gene sequencing

Amplicons w ere extracted from 2% agarose gels and purified using the Axy Prep DNA Gel Εxtraction Kit (Axygen Biosciences, Union City, CA, United States)according to the manufacturer's instructions and qualified using QuantiFluor-ST(Promega, United States). Purified amplicons were pooled in equimolar and pairedend sequenced (2 × 250) using Illumina Hiseq 2500 following standard protocols. The raw reads were deposited into the NCBI Sequence Read Archive database (Accession Number: SRP157001).

Sequencing data analysis

The raw data w ere filtered to obtain clean read s by eliminating the adapter pollution and low-quality sequences. Paired end clean read s w ere merged as raw tags using FLASH (Fast Length Adjustment of Short reads, v 1.2.11) with a minimum overlap of 10 bp and mismatch error rates of 2%[16]. Noisy sequences of raw tags were filtered by QIIMΕ (v1.9.1) pipeline under specific filtering conditions to obtain high-quality clean tags[17]. The tags were then clustered as Operational Taxonomic Unit (OTU) by scripts of USΕARCH (v 7.0.1090) softw are w ith a 97% similarity threshold[18]. The rep resentative OTU sequences w ere taxonomically classified using Ribosomal Database Project classifier v.2.2 trained on the Greengenes d atabase[19,20]. Finally, an OTU table and a phylogenetic tree were generated for diversity analysis. To estimate the d iversity of the microbial community of the sample, w e calculated the w ithinsample (alpha) d iversity by Wilcoxon rank test for two groups and multiple group comp arisons w ere mad e using Kruskal-Wallis test. Beta diversity w as estimated by computing w eighted Unifrac distance and w as visualized w ith principal coord inate analysis (PCoA). Statistical d ifferences (P ＜ 0.05) betw een the tw o group s in the relative abund ance of bacterial phyla and genera w ere evaluated using Metastats(Kruskal-Wallis test for more than two groups).

According to the genera average abundance in patients before and after treatment,genera were divided into before-enriched and after-enriched. The correlation netw ork of genera differentially enriched in before and after group was constructed by Pearson correlation test based on the abund ance. The correlation netw ork w as visualized using Cytoscape (version 3.3.0). Pearson correlation test w as also p erformed for investigating microbial d ysbiosis ind ex (MDI) and CDAI, as w ell as d ifferential genera and predicted pathways.

Functional annotation

The metagenomes of the gut microbiome w ere imputed from 16S r RNA sequences w ith PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States). This method p red icts the gene family abund ance from the phylogenetic information with an estimated accuracy of 0.8. The closed OTU table was used as the input for metagenome imputation and was first rarefied to an even sequencing depth prior to the PICRUSt analysis. Next, the resulting OTU table w as normalized by 16S r RNA gene copy number. The gene content was predicted for each individ ual. Then, the p red icted functional composition profiles w ere collap sed into level 3 of the KΕGG database pathw ays. The output file w as further analyzed using Statistical Analysis of Metagenomic Profiles (STAMP) software package[21].

RESULTS

Bacterial microbiota diversity

To determine the effect of disease activity on microbial composition, we collected four parts of mucosal samples including ileum, ascending colon, descending colon and rectum from nine patients in active and remission stages. Patients characteristics are described in Supplement Table 1. The CDAI was significantly decreased in the After group compared to the Before group, indicating that the remission had been induced after clinical treatments (Figure S1). The Before group included samples from three anatomical sites (ileum, ascending colon, descending colon) and the After group supplemented with rectum mucosa, except for three samples that were unqualified for sequencing. Given the ethical issue, we did not collect mucosa from the four sites in healthy subjects, but the samples in each site were relatively uniform (3 from ileum,4 from ascending colon, 3 from descending colon and 4 from rectum). After filtering and bioinformatic processing, a median yield of 275041 high-quality reads were obtained per sample.

The gut microbiota richness, measured by observed species and Shannon index,was not significantly different among ileum, ascending colon, descending colon and rectum in both CD and HC (Figure S2, numbers of observed species, P = 0.12 for CD and P = 0.49 for HC; Shannon index, P = 0.78 for CD and P = 0.91 for HC, Kruskal-Wallis test). The analysis of beta diversity based on the unweighted UniFrac distances showed that there was no significant difference among the four regions of intestinal tract in both CD and HC (Adonis analysis, P = 0.85 for CD, P = 0.94 for HC).Interestingly, analysis of alpha diversity as calculated by number of observed species and Chao1 index revealed that the microbial biodiversity of the patients in remission was significantly increased towards the HC compared to their active stage before treatment (Figure 1A, numbers of observed species, P ＜ 0.0001; Chao1 index, P ＜0.0001; Kruskal-Wallis test). Beta diversity represented by PCoA analysis clearly showed that the samples from patients after treatment, which clustered separately from those before treatment, tend ed to ap p roach that of HC (Figure 1B, Ad onis analysis, P = 0.001).

Bacterial microbiota composition and correlations

To investigate the specific changes of microbiota in patients w ith active and remission CD, w e assessed the relative abund ance of taxa before and after treatment. At the phylum level, the Fusobacteria w as low er in the After group than the Before group(Figure 2A). In addition, the Firmicutes was significantly decreased in CD whereas the Proteobacteria w as overrep resented in CD relative to HC. At the genus level, w e observed 65 bacterial taxa that d isplayed different abund ance betw een Before and After group. Compared w ith the After group, 14 bacterial taxa were enriched in the Before group, w hile 51 bacterial taxa were depleted in the Before group. The Beforeenriched bacterial taxa includ ed Fusobacterium, Streptococcus, Bacillaceae, etc. Bacterial taxa that w ere d ep leted in the Before group includ ed Anaerostipes, Roseburia,

Ruminococcus, Lactobacillus, Akkermansia, Lachnospira, etc, w hich w ere significantly more abundant in HC than CD (Figure 2B). Then, we used these taxa to calculate the MDI, w hich is the log of (total abund ance in organisms increased in Before group)over (total abund ance of organisms d ecreased in Before group) for the samples. The MDI show ed a positive correlation w ith clinical d isease severity (CDAI), and a negative correlation w ith species richness (Figure 2C), suggesting the close association between gut microbiota disorder and disease activity.

We further comp ared the effect of d ifferent medications includ ing thalidomid e,azathioprine (AZA), AZA p lus p red nisolone and IFX on gut microbiota. First, the observed sp ecies and Chao1 ind ex, w hich represents alp ha d iversity, tend ed to increase after treatment in four groups, although without significance in IFX and AZA(Figure S3, numbers of observed species, P ＜ 0.05 for thalid omid e, P ＜ 0.01 for AZA plus prednisolone, P ＞ 0.05 for AZA and IFX; Chao1 index, P ＜ 0.05 for thalidomide, P＜ 0.01 for AZA plus pred nisolone, P ＞ 0.05 for AZA and IFX, t test). The analysis of beta diversity based on the unweighted UniFrac d istances show ed that the overall microbial composition of the After group w as significantly different from the Before group regardless of treatment drugs (Adonis analysis, P = 0.005 for thalidomide, P =0.019 for AZA, P = 0.004 for AZA plus prednisolone, P = 0.018 for IFX). Furthermore,the d ifferential taxa betw een After and Before groups w ere id entified accord ing to each treatment. The relative abundance of Roseburia, Ruminococcus and Anaerostipes was significantly increased after IFX treatment w hile the number of Fusobacterium and Streptococcus w as low er. The relative abund ance of Roseburia, Ruminococcaceae and Lachnospira w as significantly increased w hile Fusobacteriaceae w as d ecreased after patients w ere treated w ith AZA. In the AZA plus pred nisolone group, the relative abundance of Lactobacillus, Ruminococcus and Lachnospiraceae w as increased w hile that of Fusobacterium and Bacillaceae w as decreased after treatment. In the thalidomid e group, the relative abundance of Lactobacillus and Roseburia w as also increased after treatment (Sup p lement Table 2). Collectively, these alterations in microbiota comp osition betw een Before and After group w ere similar among the four medications.

Spearman correlation test w as performed to evaluate the relationships among the genera id entified in MDI. Significant positive correlations were found in the genera d ep leted in the Before group includ ing Lachnospira, Roseburia, Anaerostipes,Ruminococcus, w hich w ere abundant in HC, suggesting their synergy as commensal bacteria in maintaining gut microbiota homeostasis and p romoting the mucosal healing process (Figure 3). On the other hand, the genera enriched in the Before group such as Fusobacterium and Bacillaceae show ed negative correlations w ith those depleted in the Before group, indicating an antagonistic relationship betw een harmful bacteria in the active stage and beneficial bacteria in the remission stage.

Microbial functions altered during CD treatment

To infer the metagenome functional content based on the microbial community profiles obtained from the 16S r RNA gene sequences, we used PICRUSt[22]. The pathw ays includ ing Lip opolysaccharide (LPS) biosynthesis proteins and LPS biosynthesis, which were enriched in patients before treatment compared to HC,tended to decrease after the induction of remission (Figures 4A and S4). On the other hand, the pathways including Folate biosynthesis, Starch and sucrose metabolism as well as Glycolysis/Gluconeogenesis w hich w ere deficient in active CD patients tended to increase after treatment and approached that of HC. Intriguingly, the abund ance of LPS biosynthesis proteins and LPS biosynthesis w ere negatively correlated with three genera including Roseburia, Ruminococcus and Lachnospira, which were more abundant in HC and patients in remission, suggesting their potential role in anti-inflammation in CD (Figure 4B-D).

Figure 1 Richness and diversity in the mucosa-associated microbiota of the patients with Crohn's disease and healthy controls before and after induction of remission. A: The number of observed species was 388.26 ± 94.22 in the Before group, 475.03 ± 96.01 in the After group and 547.85 ± 52.85 in the healthy controls (HC) group. The Chao1 index was 617.78 ± 161.04 in the Before group, 771.40 ± 146.62 in the After group and 864.08 ± 91.59 in the HC group; B: Plots shown were generated using the weighted version of the UniFrac-based Principal coordinate analysis. Samples from After group (green triangle) clustered separately from Before group (blue circle) while got close to the HC group (red square). HC: Healthy controls.

DISCUSSION

In the current study, we analyzed, using a 16S r RNA sequencing approach, the alterations of the gut mucosal microbiota in CD patients during their treatment. The comparison from the same cohort showed that the composition of mucosa-associated microbiota changed significantly after the induction of remission, with increased diversity as well as a restoration of potential beneficial bacteria. The MDI that was identified using differential microbiota between patients before and after treatment showed the correlation of microbial dysbiosis with disease activity.

Figure 2 The microbial dysbiosis index characterizes the activity of Crohn's disease. A: The composition of mucosa-associated microbiota at phylum level; B:

Accumulating evidence has demonstrated the disorder of gut microbiota in CD,and this disorder is considered as an essential factor in driving inflammation[7,23]. The gut microbial community of CD patients is characterized by reduced diversity as well as compositional changes in phylum level, including a decreased abundance of Firmicutes and an increased abundance of Proteobacteria when compared to HC[24].How ever, it remains unclear w hether these alterations of gut microbiota are associated with disease activity. Consistent with previous studies, we found the dysbiosis of mucosa-associated microbiota in patients with CD and then analyzed its comp osition before and after the ind uction of remission. Due to the d istinct microbiome signatures in different sub-phenotypes of CD, we enrolled the patients with the same behavioral phenotype to exclude bias[25]. Both the alpha and beta diversity showed that the structure of gut microbiota in patients before treatment was significantly d ifferent from those after treatment, ap p roaching that of HC and ind icating the p artial restoration of microbial balance after treatment. A previous study by Wang et al[9]reported the dynamic changes of fecal microbiota d uring IFX therapy in pediatric CD, and IFX has been demonstrated to diminish CD-associated gut microbial d ysbiosis. Another earlier stud y using a polymerase chain reactiondenaturing gradient gel electrophoresis also found that treatment w ith Adalimumab induced short-term changes in the microbiota composition, and these changes seem to parallel the partial recovery of the gut bacterial ecology[26]. AZA is the most commonly employed 6-mercaptopurines drug, and these d rugs have been found to exert antiinflammatory effects by targeting not only human macrop hages but also the gut microbiota linked to CD[27]. Consistent with these previous findings, the present study show ed that the mucosa-associated gut microbiota changed after treatment,regard less of the d rug used, and thus w e sp eculate that the alterations of gut microbiota may be associated w ith the change in d isease status from active to remission. The causal relationship betw een d isease activity and microbiota, how ever,needs further investigation using a germ-free animal model.

To further clarify the critical taxa that may be associated with the activity of CD, we id entified 65 genera show ing significant difference before and after treatment, and these genera w ere used to calculate the MDI. Interestingly, the positive correlation between MDI and CDAI indicates that the activity of CD may be associated w ith the dysbiosis of gut microbiota. A previous stud y in treatment-naïve children w ith CD also described the MDI using the differential taxa betw een CD and healthy subjects,and then d emonstrated that the MDI characterized CD severity[28]. Some MDIassociated taxa w ere common to both stud ies including Fusobacterium, Lachnospira,Ruminococcus and Parabacteroides. The relative abundance of Lachnospira, Ruminococcus and Roseburia, w hich are know n to p rod uce short chain fatty acid s (SCFAs) w as significantly increased in patients w ith clinical remission compared to those w ith active CD. SCFAs, w hich are mainly composed of acetate, propionate and butyrate,are the end products of the fermentation of dietary fiber by the gut microbiota and have been show n to exert multip le beneficial effects on mammalian energy metabolism[29]. Recently, numerous studies have show n the loss of SCFA-producing taxa in CD and that IFX treatment w as able to restore their levels in responsive patients, indicating their association with disease severity[9,24,30]. The beneficial effect of SCFAs on CD is p robably d ue to their anti-inflammation cap acity w hich is documented both in vitro and in vivo[31]. Chronic inflammation is a hallmark of CD and results from the recruitment and activation of immune cells from the circulation.Butyrate elicits anti-inflammatory effects via the inhibition of IL-12 and the upregulation of IL-10 production in human monocytes, repressing production of proinflammatory molecu les TNF-α, IL-1β, nitric oxid e, and red ucing NF-κB activation[32,33]. A clinical study explored the therapeutic effect of butyrate on patients with CD and found that the administration of butyrate induced clinical improvement and remission in 53% of p atients, in w hom butyrate successfully dow nregulated mucosal levels of NF-κB and IL-1β[34]. Thus, it is conceivable to modulate the dysbiotic gut microbiota using probiotics, prebiotics and fecal microbiota transplantation (FMT)for the management of CD. Several studies have reported the potential of FMT for CD treatment, and w e sp eculate that it may help ful to select d onors w ith a high abundance of the beneficial bacteria, w hich are deficient in patients to improve the therapeutic efficacy[35].

Analysis of the inferred metagenome in our study show ed that the abundance of LPS biosynthesis proteins and the LPS biosynthesis pathway in patients w ith CD were significantly decreased after treatment w hile the abund ance of Folate biosynthesis,glycolysis/gluconeogenesis, starch and sucrose metabolism w ere increased,suggesting that the induction of remission could partially rectify the dysbiosis of gut microbiota and restore the homeostasis of metabolic function. LPS, one imp ortant component in the outer membrane of gram-negative bacteria, plays a critical role in triggering inflammatory responses that could further result in various diseases such as CD[36]. The serum levels of LPS w ere demonstrated to increased markedly in active CD p atients comp ared w ith p atients in remission and HC and w ere p ositively correlated w ith the severity of the d isease[37]. Moreover, the blockade of intestinal mucosal inflammation w ith IFX could red uce the levels of LPS and IFX has been rep orted to d iminish the CD-associated gut microbial d ysbiosis[9,37]. Analysis of microbiota genes in this study show ed a prominent upregulation of LPS-related pathw ays in patients before treatment, and this upregulation may be due to the overgrowth of LPS-producing bacteria such as Fusobacterium and Eikenella. Folate deficiency is common in patients w ith IBD, mostly in CD patients. Up to 80% of patients with CD present with low levels of serum folate, and this is more common in active disease than at times of remission[38,39]. Folate is produced in large quantities by the colonic microbiota, mainly probiotics such as Lactobacillus and Bifidobacterium[40].The enrichment of Lactobacillus in CD patients after treatment as shown in our study may promote the biosynthesis of folate whose deficiency is associated with disease severity.

Figure 3 Spearman correlations among the 65 Crohn's disease-associated genera in gut mucosal samples. The green circle represented taxa enriched in patients after treatment while the red circle represented taxa enriched in patients before treatment. Lines between nodes denoted Pearson correlation (r ＞ 0.2 and P ＜0.05). The red and blue lines represented the positive and negative correlation, respectively.

This study has some limitations. First, the number of patients was somewhat small.Therefore, the characteristics of mucosa-associated gut microbiota d uring CD treatment and the predictive value of MDI in disease activity warrant verification in large populations. Second, the patients in this study received different drugs for inducing remission, w hich may cause some bias. Our results showed that the composition of gut microbiota in patients before treatment was significantly different from those after treatment regardless of the drug used. Due to the small sample size in each treatment group, the effect of d rugs on gut microbiota needs further investigation.

In conclusion, the results in this study presented that both the structure and function of mucosa-associated bacterial microbiota in patients w ith CD changed significantly during treatment, approaching healthy status. The dysbiosis of gut microbiota, which is associated with disease activity, is partially restored after the induction of remission with characteristics including increased biodiversity and an increase in SCFA-producing bacteria. Therefore, the disturbance of gut microbiota,especially the overgrowth of pathogenic bacteria and the depletion of beneficial bacteria may act as a potential therapeutic target for CD treatment.

Figure 4 The predicted functional module involving pro-inflammatory pathways altered in Crohn's disease compared to healthy controls. A: Pathways including Lipopolysaccharide biosynthesis proteins and Lipopolysaccharide biosynthesis predicted to show significant different abundances among before, after and healthy controls group according to Kyoto Encyclopedia of Genes and Genome pathway analysis. The Crohn’s disease-depleted genera including Roseburia,Ruminococcus and Lachnospira were negatively correlated with Lipopolysaccharide biosynthesis proteins (P = 0.001 for Roseburia, P = 0.002 for Ruminococcus and P = 0.025 for Lachnospira) and Lipopolysaccharide biosynthesis (P = 0.0002 for Roseburia, P = 0.002 for Ruminococcus and P = 0.021 for Lachnospira). B:Roseburia; C: Ruminococcus; D: Lachnospira.

ARTICLE HIGHLIGHTS

Research background

Accumulating evidence demonstrated the alterations of gut microbiota in patients w ith Crohn's d isease (CD) compared to healthy subjects. How ever, comparative analysis of mucosal microbiota in the same cohort of patients before and after treatment remains limited. The different characteristics of mucosa-associated gut microbiota betw een active and quiescent CD may provide as a predictive tool for disease relapse as w ell as a potential therapeutic target for treatment.

Research motivation

Most studies investigating the gut microbiota of CD have used fecal samples while only few studies have investigated the mucosal microbiota, which is believed to directly affect epithelial function and may be more deeply involved in the pathogenesis of CD. Although the dysbiosis of gut microbiota have been reported in patients with CD as compared with healthy controls, the microbial changes during treatment and their association w ith disease activity are largely unknown.

Research objectives

To illustrate the global alterations of mucosa-associated microbiota in patients with active CD before and after treatment.

Research methods

A total of 74 mucosal biopsies w ere collected from 15 participants including 9 patients w ith CD and 6 healthy individuals. Sampling included both active and remission stages for each patient w ho underw ent clinical treatment. The gut microbiota was sequenced by 16S r RNA analysis.

Research results

Our results show ed that the structure of gut microbiota in patients w ith active CD changed significantly after the induction of remission, including the decreased abundance of pathogenic bacteria and increased abundance of beneficial bacteria.

Research conclusions

The dysbiosis of mucosa-associated gut microbiota in active CD w as partially restored after treatment, indicating the association of microbiota and disease activity.

Research perspectives

The variations of gut microbiota may act as a tool to supervise and predict the recurrence of CD,and the maintenance of microbial homeostasis could become a potential therapeutic target for the disease.