The microbiome of Crohn’s disease aphthous ulcers

Crohn’s disease (Crohn’s disease), a chronic, relapsing inflammatory disease of the gastrointestinal tract, is thought to result from an aberrant, ongoing immune response to bacteria, in genetically susceptible individuals. Over 200 gene variants are associated with IBD, with just over 30 of these being Crohn’s disease-specific [1]. Analysis of these gene variants suggests that host–microbe interactions are crucial in the development of Crohn’s disease.

Several lines of evidence suggest that microbes play a role in either the onset or perpetuation of Crohn’s disease. The earliest lesion in Crohn’s disease is the aphthous ulcer, which overlies Peyer’s patches in the small bowel, and lymphoid follicles in the large bowel. These lymphoid aggregates are the site of luminal antigen sampling by innate immune cells. Granulomas, which are a histological hallmark of Crohn’s disease, contain bacteria [2]. Temporary diversion of the fecal stream to a proximal ileostomy prevents the recurrence of inflammation in down-stream mucosal sites [3]. Numerous studies have shown that the gut microbiome is dysregulated in Crohn’s disease, both in terms of its species composition and its function [4‐6].

Reduced alpha diversity (mean number of bacterial species in a given sample) is frequently observed in the microbial communities of Crohn’s disease mucosa when compared to mucosa from healthy controls, and cannot be attributed to inter-individual variation in the gut microbiome [7]. Studies consistently show that the gut microbiome of Crohn’s disease patients is depauperate, in particular butyrate-producing Faecalibacterium prausnitzii and Roseburia [8‐11]. Other groups of bacteria, such as Enterobacteriaceae, which includes Escherichia coli, are increased in Crohn’s disease relative to controls [4, 12‐14], and their abundance has been shown to correlate with disease severity [4]. Although a number of bacterial taxa have been implicated in Crohn’s disease, no single causative organism has been identified.

There is tremendous inter-individual variation in the gut microbiome of healthy individuals; however, despite this, the functional capacity of each individual’s microbiome remains similar [7]. Conversely, modest differences in the taxonomic composition of the gut microbiome of patients with IBD are associated with major changes in its function [15]. These changes may reflect the response of bacteria to an inflamed gut, as enrichment in microbial pathways that enable bacteria to cope with oxidative stress, evade immune responses, and take up host metabolites without prior synthesis (auxotrophy) is observed. There are also corresponding reductions in short chain fatty acid (SCFA) and amino acid biosynthesis, as well as gut carbohydrate metabolism [15]. Proving that changes in the gut microbiome precede the onset of disease, cause a disease flare, or are a consequence of inflammation remains challenging.

It is almost impossible to obtain mucosal samples from people prior to the development of IBD, and only a few studies have assessed the mucosal microbiome of patients with Crohn’s disease at the onset of disease. A large study of paediatric patients with new-onset Crohn’s disease showed that species-richness was reduced in Crohn’s disease, and that the abundance of several taxa was altered [4]. However, an inflammatory response was well established in 96% of the patients recruited to this study [mild (PCDAI 10–30)—moderate/severe disease (PCDAI > 30)]. Dysbiosis is associated with other inflammatory conditions, such as obesity [16] and Type 2 diabetes [17], suggesting that chronic inflammation drives changes in the gut microbiome. A study by Kiely et al. showed that the mucosal microbiome of patients with inflammatory bowel disease (IBD) fluctuates over time, with greater changes observed in patients who had ongoing microscopic inflammation [18]. The majority of studies support the idea that dysbiosis is a common response to chronic inflammation.

The earliest mucosal lesions in Crohn’s disease, aphthous ulcers, are small (1–5 mm) superficial ulcerations surrounded by a ring of erythema then normal surrounding mucosa [19]. These lesions overlie the follicle associated epithelium (FAE) of the small bowel (Peyer’s patches) and large bowel (lymphoid follicles) [20], can be found in 70% of patients with Crohn’s disease [21], appear more commonly in the distal ileum [19, 21], and can develop into larger, transverse linear ulcers [19]. Approximately 10% of the epithelial cells of the FAE are microfold cells, commonly referred to as ‘M’ cells. These cells have a reduced glycocalyx and blunted microvilli, and are highly specialized in phagocytosis and transcytosis of luminal antigens, which they package in vesicles and deliver to underlying immune cells [22].

Numerous bacterial pathogens, including Yersinia pseudotuberculosis [23], Mycobacterium tuberculosis [24], Salmonella typhimurium [25], Shigella flexneri [26], and Escherichia coli [27, 28] exploit M cells to cross the epithelial barrier and cause infection. Adherent invasive E. coli (AIEC), which have been implicated in Crohn’s disease, target M cells on Peyer’s patches through the expression of one of two major long polar fimbriae (lpfA) operons, which allows them to translocate the intestinal epithelial barrier [29]. However not all AIEC strains harbor lpfA [30], and non-AIEC strains may harbor lpfA, suggesting that this is not the only mechanism by which AIEC exploit M cells of the FAE.

Because of the difficulty in obtaining biopsy samples before Crohn’s disease presentation, aphthous ulcers represent the earliest stage at which microbial communities can be assessed at the onset of disease or a disease flare [31]. The aim of our study was to compare the microbial communities of aphthous ulcers and adjacent mucosa from individuals with Crohn’s disease with mucosa from healthy controls, to determine whether or not specific bacteria, or an imbalance in the gut microbiome, are present in the initial Crohn’s disease lesion. This is the first study to assess the bacterial community composition of aphthous ulcers in Crohn’s disease.

This is the first study to assess the microbiome of aphthous ulcers in Crohn’s disease. Our study suggests that the microbiome is not imbalanced in the initial Crohn’s disease lesion, relative to control mucosa. The alpha diversity, and composition of the microbiome of aphthous ulcers and adjacent mucosa from patients with IBD was similar to mucosa from controls. We found no evidence for a reduction in the genus, Faecalibacterium, which only contains one species, Faecalibacterium prausnitzii, and is commonly found to be decreased in Crohn’s disease. We did not detect an increase in taxa that are usually over-represented in Crohn’s disease mucosa, such as the family Enterobacteriaceae, which includes E. coli.

Bacterial community imbalance, or dysbiosis, is a common finding in IBD. Studies often report a decrease in protective groups (such as Lachnospiraceae, Roseburia, and Faecalibacterium), and a subsequent increase in pathobionts, (such as Proteobacteria, Ruminococcus, and Fusobacterium). Dysbiosis is likely to result from several factors. One study looked at the effects of inflammation, antibiotic exposure, and diet (exclusive enteral nutrition [EEN]) on the gut microbiome of paediatric patients with active Crohn’s disease [33]. They found that each factor independently affected different bacterial taxa in the microbial community. They also showed that dysbiosis decreased with reduced intestinal inflammation, and that the microbiome of patients who responded to anti-TNF therapy and EEN became more similar to healthy controls than that of non-responders. These data support the idea that dysbiosis is a consequence, not cause, of inflammation. We did not control for diet in this study, however we did observe dysbiosis in patients who had consumed antibiotics. The degree or duration of inflammation in the aphthous ulcers may not have been great enough to affect the microenvironment, or to initiate dysbiosis.

A study by Lupp et al. [34]. showed that host-mediated inflammation in response to infection (Campylobacter jejuni) and oral administration of dextran sodium sulfate (DSS) leads to dysbiosis in a mouse model. In particular, they observed an expansion in Enterobacteriaceae. A study of the microbiome of a cohort of treatment-naïve new-onset patients with Crohn’s disease, revealed that antibiotic use exaggerates dysbiosis [4]. They also showed that inflammatory conditions were strongly associated with a reduction in species richness and expansion of Enterobacteriaceae, as well as Bacteroidales, and Clostridiales.

The strength of this study was the ability to assess the microbiome of the initial Crohn’s disease lesion (aphthous ulcer), before transmural inflammation and clinical manifestations developed for the first time, or for a new flare. Although our Crohn’s disease cohort was small, we were able to demonstrate that dysbiosis is not a feature of aphthous ulcers. Similar sized cohorts of patients have demonstrated dysbiosis in samples obtained from patients with active Crohn’s disease, including a reduction in Faecalibacterium [35, 36]. If dysbiosis were a feature of the aphthous ulcer microbiome, we would likely have observed it in a number of our Crohn’s disease patients.

It is unclear when dysbiosis of the gut microbiome develops in patients with Crohn’s disease, and if it becomes progressively worse with each disease flare. One study assessed the gut microbiome of unaffected genetically-linked relatives of children with Crohn’s disease. The unaffected relatives had alterations in their gut microbiome in the direction of their relatives with Crohn’s disease, but did not display a distinct dysbiosis [36]. The findings of this study suggest that dysbiosis develops close to disease onset, or as a consequence of the disease process. There is some evidence to suggest that dysbiosis improves over time, but is still evident, in patients with complete mucosal healing or who have responded to treatment [35]. Only one patient with Crohn’s disease in this study (CD11) had previous surgery, all other patients were in remission, or had only mild symptoms. Patients with a long history of mild disease may be less likely to have gut microbiome imbalance.

If dysbiosis does improve in the absence of active disease, then interventions aimed at restoring the gut microbiome may be effective in increasing gut microbial diversity. Reduced alpha diversity could lead to a break-down in the functional redundancy of gut communities, which may exacerbate symptoms. It would be important to administer interventions, such as pro-, pre- and syn-biotics, in the absence of inflammation, as attempts to establish or nourish bacteria that do not cope well in an environment of chronic inflammation and oxidative stress may be futile.

Our data suggest that dysbiosis is a consequence of the inflammatory disease process, as it was not observed in the initial lesion. We did not detect dysbiosis in the three patients who were newly diagnosed at the time of sampling, nor in patients with more established disease. Longitudinal studies aimed at assessing the gut microbiome before disease onset and throughout successive flares would provide further insight into the nature and development of dysbiosis, but acquiring samples from patients prior to diagnosis remains problematic.