|Home | About | Journals | Submit | Contact Us | Français|
Mucosa-associated bacterial flora from control individuals and inflammatory bowel disease (IBD) patients were evaluated by real-time analysis using 16S rRNA-based genus-specific primers. Our data show a clear delineation in concentration of bacteria between the predominating and subdominating genera under disease conditions, indicating that the subsets of bacteria participating in the pathogenesis of ulcerative colitis (UC) and Crohn's disease (CD) are likely to be different.
Our gut microbiota plays a very important role in protection of gut mucosa (8, 14), and thus, dysbiosis may contribute to processes of mucosal injury (17, 18, 19). The microbiota close to the mucosa, which differs from the luminal microbiota, has so far received less attention, and yet the mucosa-associated microbiota is very close to the inflammatory process. As inflammatory bowel disease (IBD) is a disorder of mucosal inflammation, the mucosa-associated microflora seems to be of great relevance to the disease process (1, 20). Differences have been observed between the dominant fecal microbiota and the mucosa-associated microbiota at different sites of the colon and rectum in IBD versus healthy subjects (7).
To date, no specific bacterial agents have been identified as potential factors triggering intestinal inflammation, although the involvement of pathogenic bacteria cannot be excluded. Anaerobic bacteria (Escherichia coli mpk) have been shown to induce colitis, whereas a protective role for Bacteroides vulgatus mpk has been shown to occur in several IBD animal models (23). Decreases in the relative proportions of phylogenetic groups (Bacteroides, Eubacterium, and Lactobacillus spp.) have been reported to occur during the active and remission stages of ulcerative colitis (UC) cases (11).
Our study included 84 patients (72 with UC and 12 with Crohn's disease [CD]) and 65 individuals with noninflammatory gastrointestinal conditions as healthy controls, recruited by the Department of Gastroenterology of the All India Institute of Medical Sciences, New Delhi, India. Diagnosis was based on a combination of clinical history, colonoscopic findings, endoscopic imaging, and histological features. The clinical characteristics of the patients are given in Table Table1.1. The inclusion criteria allowed for concurrent medication with oral 5-aminosalicylates (5-ASA) at a stable dose for at least 4 weeks or azathioprine-6-mercaptopurine at a stable dose for at least 3 months prior to screening. Colon preparation was done with polyethylene glycol (Peglec) dissolved in water and administered orally on the day of the procedure. The biopsy specimens were collected from the rectosigmoid area of the colon for all patients and from the normal sigmoid mucosa for the control group. The control population consisted of patients who were taken up for sigmoidoscopy for evaluation of either irritable bowel syndrome or anal hemorrhoids. This study was approved by the ethics committee of the All India Institute. The biopsy samples, with associated mucus, were collected aseptically, rinsed gently with sterile anaerobic phosphate-buffered saline, and frozen immediately at −80°C for subsequent PCR analysis. Total DNA from the colon biopsy samples (0.1 g) was extracted according to a modified version of the protocol of Taggart et al., and DNA was finally eluted in 50 μl of Tris-EDTA (TE) buffer (21). One microliter (20 ng) of this DNA from each sample was used to analyze the bacterial population. Primer sets were designed to differentiate major anaerobic genera from the 16S rRNA region (see Table S1 in the supplemental material). The dissociation curves of the primers that were designed by us are represented in Fig. S1 in the supplemental material. Reference clones were raised from genomic DNA of healthy individuals by use of genus-specific primers, sequenced, and deposited into the EMBL database (see Table S2 in the supplemental material).
In real-time analysis, the standard curve was constructed by serial dilutions of each reference clone prepared from 0.05 to 50,0000 pg/tube, corresponding to 1 × 101 to 1 × 107 copy numbers. The standard curve (see Fig. S2 in the supplemental material) of the reference clones was used to extrapolate the numbers of bacteria in the biopsy specimens. With the molecular mass of the plasmid and insert known, it is possible to calculate the copy number as follows (26): mass in Daltons (g/mol) = (size of double-stranded [ds] product in base pairs [bp]) (330 Da × 2 nucleotides [nt]/bp). Hence, the g/mol value divided by Avogadro's number equals the g/molecule value, which equals the copy number.
For calculation of significant differences in mean bacterial population (means of results from three experiments with each sample), the Student t test was used, with P values of 0.05 or below considered significant. The real-time data represented in Fig. Fig.11 depict the average concentration of each bacterial genus observed in control individuals versus UC and CD patients. Among the predominating genera, the concentrations of bacteria followed the pattern Bacteroides > Bifidobacterium > Ruminococcus > Eubacteria > Clostridium leptum subgroup, and among the subdominant group, the pattern was Campylobacter > sulfate-reducing bacteria (SRB) > Lactobacillus > Methanobrevibacter > Peptostreptococcus productus. The Bacteroides group was abundant in healthy samples; however, there were significant drops in its concentrations in UC (P = 0.0108) and CD (P = 0.0088) patients. We observed significant decreases in the populations of Lactobacillus, Ruminococcus, and Bifidobacterium bacteria in both UC (P = 0.0206, 0.0094, and 0.031, respectively) and CD (P = 0.016, 0.0036, and 0.035, respectively) patients. The losses of normal intestinal bacterial taxa, such as Bacteroides, Lactobacillus, Ruminococcus, and Bifidobacterium, observed by us supported earlier observations (3, 11, 16, 27) proving the hypothesis that loss of commensal organisms profoundly modifies gut mucosal homoeostasis through loss of essential micronutrients (short chain fatty acids) and redox potential (2). The Clostridium leptum subgroup was reduced nonsignificantly in CD patients but increased sharply in UC patients (P = 0.0155), as observed earlier (9, 18, 27). The difference was more pronounced between UC and CD patients (P = 0.0118). The C. leptum group encompasses several butyrate-producing bacterial strains; hence, the decreases in their populations could play a role in the onset of CD. That there were increases in coccoid rods of Gram-positive Eubacterium (P = 0.007) and Peptostreptococcus (P = 0.017) bacteria in CD patients but not in UC patients supported earlier observations indicating that the roles of the subsets of bacteria in the pathogeneses of UC and CD may be different (22). The populations of Campylobacter bacteria exhibited significant increases in UC (P = 0.0113) as well as in CD (P = 0.0174) patients when compared with the levels for the controls, and the levels for UC and CD patients were significantly different from each other (P = 0.0133). With the severity of the disease, the Campylobacter population increased significantly but reverted to normal during the remission stage (P = 0.0083), as calculated by the Mann-Whitney U test (Fig. (Fig.2).2). Such a pattern was not observed in other genera under similar conditions (data not shown). Our study also recorded increases in the populations of two subdominant inhabitants, the methanogenic bacterium Methanobrevibacter (P = 0.0069 for UC patients and 0.017 for CD patients) and sulfate-reducing bacteria (SRB) (P = 0.0491 for UC patients and 0.0287 for CD patients), compared with the levels for the controls. Earlier reports indicate higher levels of breath methane in UC and colonic polyposis patients than in controls and detected higher rates of methanogenic bacteria from fecal specimens of patients than from those of controls (12, 25). That there were higher numbers of SRB and higher rates of sulfidogenesis in UC patients than in controls supported our findings (4, 5). The colonic mucosa colonized by SRB and showing simultaneous production of H2S has the potential to cause colonic injury in UC patients (13, 19). This result has been linked to the pathophysiology of some chronic colonic disorders (10). The above-mentioned findings indicate either a predominance of some potentially harmful bacterial groups or a decrease in beneficial bacterial species in the mucosal flora of IBD patients. Our results demonstrate a method for calculating bacterial populations accurately from real-time data. Our findings suggested that a differential change in a subset of a bacterial population may be associated with UC and CD. This suggests a testable hypothesis: alterations in the representation of components of the microbiota described here may be biomarkers that may help to predict disease predisposition, activity, severity, and responsiveness to therapy.
This study was supported by a research grant to J.P. from the Department of Science and Technology, Government of India. R.V. is grateful to the University Grants Commission, New Delhi, India, and A.K.V. to the Indian Council of Medical Research, New Delhi, India, for Research fellowships.
We gratefully acknowledge the subjects who participated in this study.
Published ahead of print on 22 September 2010.
†Supplemental material for this article may be found at http://jcm.asm.org/.