Final Progress Report
Proposal No. IBD-0032
Principal Investigator: Timothy Florin, MBBS
Applicant Organization: The University of Queensland (Brisbane, Australia)
Project Title: Mucosal bacterial pathogens in inflammatory bowel disease
Period of Award: December 1, 2002 – February 28, 2005
1. The project aims were to:
a. describe the mucosal bacterial community in control and inflammatory bowel disease (IBD) - Crohn’s disease (CD) and ulcerative colitis (UC) - subjects by using ‘random’ cloning of 16S rRNA bacterial sequences. This methodology gives a different perspective to the identification by traditional culture methods of the mucosal bacterial community.
b. use the phylogenetic analysis of the clone libraries to discern mucosal bacteria that were associated with IBD. The phylogenetic analysis is based on a bacterial ribosomal database called ARB.
c. test these associations by using quantitative means to measure these bacteria in mucosal biopsies. We used real time PCR.
d. culture IBD-associated bacteria.
2. Accomplishments towards meeting those aims:
a. We initially made clone libraries of partial length 16S rRNA genes, but realized that bacterial identification of the partial length sequences by phylogenetic analysis using the available ribosomal database was not as accurate as when using full length 16S rDNA sequences.
Therefore, we made clone libraries of full length 16S rDNA sequences and analyzed them in our ARB database. This, of course, required considerably more resourcing. The full length sequence clone libraries were from:
- four ulcerative colitis patients (from four inflamed distal colonic + four non-inflamed proximal colonic sites) x two = eight clone libraries
- six Crohn’s distal colonic non-inflamed sites
- six non-IBD healthy control subjects matched for age and sex, colonoscoped as part of colon cancer screening
b. There were over 180 different bacterial 16S rDNA sequences from all libraries, with 120 being represented at least twice in the clone libraries. Ninety-nine percent of the sequences were from Firmicutes (48%), Bacteroidetes (33%) and Proteobacteria (18%) phyla. The most common bacterial sequence was identical to Bacteroides vulgatus 16S ribosome, but 73% of sequences were not from known or previously cultured bacteria. The former agrees nicely with traditional culture methodology. The latter could never be appreciated by the culture techniques. There were no statistically significant differences between controls or patient groups, but some sequences appeared to be more common in IBD (P < 0.05, when performed without Bonferroni correction for multiple comparisons).
One of these 16S rDNA sequences was identical to that of CDAB, which has been isolated in pure culture. Three others, also in the Firmicutes phylum, were not related to any known previously cultured bacteria.
We found that sequences identical with Bacteroides vulgatus, appeared less frequently in the clone libraries of inflamed and non-inflamed UC.
c. The clinical significance of the findings using the semi-quantitative random PCR methodology and the associated P values based on univariate parametric statistics, required confirmation/refutation by a quantitative method. We have so far restricted this part of our endeavor to the known culturable bacteria, CDAB and Bacteroides vulgatus. We now have novel quantitative data on a larger second biopsy set using real time PCR for mucosal CDAB and Bacteroides vulgatus in pinch biopsies from inflamed and non-inflamed mucosa. The IBD biopsies were from ileum, proximal and distal colon of 26 CD {Vienna classification: 7 ileal, 15 colonic, 4 ileocolonic patients} (0 on salazopyrin, 5 no treatment, 16 on thiopurines, 8 on 5-aminosalicylates, 3 both) and 20 distal UC (5 on salazopyrin, 5 no treatment, 4 on thiopurines, 9 on 5-aminosalicylates). The non-inflamed GI control subjects (NC) biopsies were from ≥ two sites and comprising all-comers matched for age and sex (14 irritable bowel syndrome patients, 6 cancer screening). Mean biopsy size as determined by the human GAPDH gene was not different between UC, CD and NC.
We used total bacteria as our internal standard. This was measured by real time PCR with the universal domain Bacteria 27f and 1492r primers. This allowed us to put a figure on the increased number of total mucosal bacteria, which others have reported increased in IBD by using semi-quantitative FISH. The first table below shows total bacteria ratios measured by real time PCR. It can be seen that mucosal bacteria were increased especially in non-inflamed IBD mucosa.
Table
From the second table below, it can be seen that CDAB is increased ~10 x in non-inflamed CD & UC, 6 x in inflamed mucosa (independent of sites x 2 = colon or terminal ileum; disease activity x 3 = quiescent, mild, severe; Rx x 3 = thiopurines, salazopyrin or no treatment/ 5-aminosalicylate: data not shown in table. (Data were log-transformed and analyzed by 2-sided t-test).
Table
While the ratios were 6 – 10 x, there was considerable overlap between NC, CD and UC. This was not due to biopsy size, as it was the same when normalized to GAPDH. Thus, the finding of increased CDAB cannot be used as a diagnostic test. The figure below illustrates this for non-inflamed colonic mucosa.
Bacteroides vulgatus was not less (P > 0.33) in non-inflamed UC and non-inflamed CD but was 50% less in inflamed UC mucosa (P = 0.02). This is interpreted as consistent with primary and secondary mucosal defense – mucin, defensins, cryptidins, IgA, cellular immunity – working to limit the bacteria
d. We have cultured CDAB, bifidobacteria bifidum, B. vulgatus and measured their ribosomal content to calculate total of each bacteria per biopsy. We plan to test the CDAB in animal models and in vitro experiments
There are different bacteria suggested by others to be involved in IBD pathogenesis. They are Desulfovibrio, bifidobacteria, F. varium, (NB: adherent intestinal E. coli is only increased in inflamed ileum, Darfeuille-Michaud A et al. Gastro 2004; 127:412-21). With the exception of Desulfovibrio now shown to be not increased in UC, there are no reliable published quantitative real time PCR data for them. In order to put our finding with CDAB in context, we have designed and validated primers for each of bifidobacteria and F. varium. To date, we have found Ct PCR data to be not different to our NC controls. They are at very low levels in the case of F. varium, higher levels of F. ulcerans (publication in preparation). (We do not hold with the recently published real time data of Okhusa et al. (Nomura T et al. Aliment Pharm Ther 2005;21:1017-27) because their primers were not specific for F. varium including human sequences and other bacteria, and detection levels were extraordinarily low (50 PCR cycles). Thus, our quantitative finding with CDAB and IBD appears to be unique. We are looking to do real time PCR on phylogenetically closely related bacteria, because we do not exclude the possibility that this may be a functional rather than bacterium-specific effect.
3. Lay summary of the report - significance:
In summary, we have novel quantitative data about human mucosal gut bacteria. We believe that the finding of increased CDAB may have implications for IBD pathogenesis. CDAB or functionally related bacteria may damage the barrier function of the intestinal epithelium. (The rationale for this belief is not detailed here and it requires experimental confirmation). We believe that the alteration in CDAB is an early event in pathogenesis because it affects non-inflamed IBD mucosa.