Inflammatory bowel diseases (IBD), including Crohn disease (CD) and ulcerative colitis (UC), are disorders that affect >4 million people worldwide (1). The incidence of IBD has significantly increased globally in the 20th century and recent observations indicate their ongoing emergence (2,3). The diagnosis of IBD is based on clinical history, physical examination, serology, radiological studies, endoscopy, and histology (4). Although no ideal noninvasive biomarker exists for IBD, the assessment of C-reactive protein (CRP) is commonly employed during the diagnostic process. CRP serum levels have shown value in indicating prognosis, response to therapy, and disease activity in CD (5–7).
About 25% of patients with IBD present during childhood or adolescence (8) with somewhat different disease characteristics than adults (9). The distinction between CD and UC in as many as 15% of adult and 20% of pediatric IBD cases can be challenging (10). The occurrence of a granuloma is pathognomic of CD and is helpful in distinguishing Crohn colitis from UC. The cause of noncaseating granulomas in CD is unknown and their presence has not been conclusively linked to specific microorganisms, including Mycobacterium avium subspecies paratuberculosis (MAP) (11,12). In the meantime, the intestinal microbiota plays a significant role in the pathogenesis of CD in which disease phenotype as well as host genotype is associated with specific shifts in composition (13) depending on the location (ileum, colon) and sample (feces, mucosa) examined (14,15). The fecal microbiota differs from mucosal (16) with the latter likely being more relevant for intestinal immunomodulation (17).
The advancement of culture-independent methodologies has shed light on the complexity and dynamic structure of the human microbiota; however, variation between interrogation methods, sampling, and the IBD populations studied has confounded the clear characterization of disease-associated microbiota and its potential pathogenic role (18). Moreover, the possible confounding effects of long-standing inflammation and ongoing anti-inflammatory treatment have rarely been accounted for in microbiota studies of IBD. In the meantime, findings demonstrated the significant microbiota-modifying effects of frequently used pharmacotherapy in the disease group (19).
Pediatric IBD cases present a unique opportunity to examine the biological components of IBD pathogenesis following a suspected shorter duration of the disease than in adults. Furthermore, if samples are obtained at the time of the diagnostic procedure, subjects will most likely be treatment naïve. The largest microbiota investigation on pediatric colonic mucosal samples from patients with IBD with limited medication exposure was conducted on 12 CD cases and 17 controls with bacterial cultures and real-time polymerase chain reaction (PCR) of the bacterial 16S rRNA gene for 14 separate species (20). Although yeasts are acknowledged as potential pathogens in CD (21), high-throughput fungal metagenomic analyses have not been applied to pediatric cases of IBD.
In the present study, we examined transverse colonic mucosal biopsy samples by massively parallel pyrosequencing of the bacterial 16S rRNA gene and the fungal small subunit (SSU) ribosomal region in 15 treatment-naïve pediatric CD cases and 26 controls. The microbiota associations of CD were correlated with histological, clinical, and laboratory characteristics of the patients.
Patients and Samples
Control patients (abdominal pain: 8, irritable bowel syndrome: 4, hematochezia: 3, solitary juvenile polyp: 3, diarrhea: 2, perianal fissure: 2, gastroesophageal reflux disease: 2, and 1 of each: gastritis, healthy [polyposis in sister]) were recruited before endoscopy following informed consent through the institutional review board approved tissue banks of the Charles University, (EK-1796/08); the Pediatric Inflammatory Bowel Disease Consortium Registry at the Baylor College of Medicine (H-17654); and the Massachusetts General Hospital (MGH; 2009p001287) (Table 1). Patients only with grossly and histologically normal mucosa at colonoscopy were designated as controls.
Treatment-naïve CD cases were recruited before their first diagnostic colonoscopy in Baylor and MGH whose disease was determined based upon clinical, biochemical, and histological characteristics. Neither patients with CD nor controls reported the use of antibiotics within 6 months of sampling. Transverse colonic mucosal samples were snap-frozen on dry ice or in liquid N2 immediately after biopsy and stored at −80°C until further analysis. Table 1 shows demographic characteristics and Montreal classification of the patients studied. Age, sex, ethnicity, and disease location were similar between granuloma-positive and -negative patients. Age and sex did not differ significantly between CD and controls. Studies indicate that the proximal and distal colonic microbiota may differ (22). Therefore, we aimed to interrogate the middle/average of the mucosal bacterial community by studying transverse colonic mucosa samples.
CRP correlation was extended to 86 treatment-naïve pediatric CD cases. Granulomatous CD was defined as the presence of granuloma or a distinct giant cell in at least 1 biopsy specimen. Histological severity of inflammation was graded between 0 and 3 (none to severe) based on the pathology reports incorporating epithelial damage, architectural distortion, and white blood cell infiltration of the lamina propria and epithelium by a physician observer (S.M.). Only those patients who had both esophago-gastroduodensocopy and colonoscopy evaluation were included. Only CRP values obtained 1 week before or at endoscopy were considered. Histological assessment of terminal ileum was made at similar frequency between granulomatous and nongranulomatous CD groups.
After thawing, the colonic mucosal biopsies were centrifuged at 14,000 rpm for 30 seconds and resuspended in 500 μL of RLT buffer (Qiagen, Valencia, CA) (with β-mercaptoethanol). Sterile 5 mm steel beads (Qiagen) and 500 μL of sterile 0.1-mm glass beads (Scientific Industries, New York) were added for complete bacterial lyses in a Qiagen TissueLyser (Qiagen), run at 30 Hz for 5 minutes. Samples were centrifuged briefly, and 350 μL of RTL and 200 μL of 100% ethanol were added to a 100-μL aliquot of the sample supernatant. This mixture was added to a DNA spin column, and DNA recovery protocols were followed as instructed in the QIAamp DNA Mini Kit (Qiagen) starting at step 5 of the tissue protocol. DNA was eluted from the column with 30 μL of water and samples were diluted accordingly to a final concentration of 20 ng/μL. DNA samples were quantified using a Nanodrop spectrophotometer (Nyxor Biotech, Paris, France).
Massively Parallel bTEFAP
Bacterial tag-encoded FLX-Titanium amplicon pyrosequencing (bTEFAP) was performed as described previously (24). (For details, see supplementary Methods, http://links.lww.com/MPG/A115.)
Bacterial Diversity Data Analysis
Before analysis, sequences shorter than 300 bp were removed and sequences were depleted of any nonbacterial ribosome sequences and chimeras using the Black Box Chimera Check software B2C2 (described and freely available at http://www.researchandtesting.com/B2C2.html). To determine the identity of bacteria in the remaining sequences, they were first queried using a distributed BLASTn.NET algorithm (25) against a database of high-quality 16S rRNA bacterial sequences derived from NCBI. (For details, see supplementary Methods, http://links.lww.com/MPG/A115.)
Fungal Diversity Analysis
Multivariate analysis of bacteria and fungi were based on measures of multivariate distance (ie, dbRDA); however, the specific distance measure most used for bacteria (UniFrac) was not available for fungi. Therefore, the Bray-Curtis distance measure was used in this case, which is a robust measure of distance for community datasets (26). The primer set to detect fungal SSU ribosomal region was as follows: SSU-F: 5′-TGG AGG GCA AGT CTG GTG-3′; SSU-R: 5′-TCG GCA TAG TTT ATG GTT AAG-3′. Barcoding, quality filtration, and taxonomic assignment were performed as in the bacterial analyses.
The DNA extraction steps for MAP samples were the same as for the bacterial diversity testing. The only difference was that we used Qiagen TissueLyser to run at 30 Hz for 20 minutes to break MAP cells. The designed MAP primers and probe were as follows: MAPIS-F: 5′-TGG GTT GAT CTG GAC AAT GAC GGT-3′; MAPIS-R: 5′-TAA CCA TGC AGT AAT GGT CGG CCT-3′; MAPIS-probe: 5′-/56-FAM/TAC GGA GGT GGT TGT GGC ACA A/3IABkFQ/-3′. MAP ATCC strains (19689, 19851, 43544, 15769, 19074, 25219) were used as positive controls, and Mycobacterium intracellulare ATCC strains (13950, 16985, 25122) were used to test primer cross-reaction. The quantitative real-time PCRs were performed by a Roche Light Cycler 480. The PCR reactions were as follows: deactivation, 95°C for 10 seconds; amplification, 35 cycles of 95°C for 15 seconds, 60°C for 1 minute (27).
Further Statistical and Bioinformatic Analysis
Unpaired, 2-tailed t-tests, odds ratio (OR) and 2-sided Fisher exact test calculations were also used in the group comparisons. Statistical significance was declared at P < 0.05. Error bars represent standard error of the mean (SEM).
Crohn Disease Associates With Significant Bacterial Microbiota Separation
Distance-based redundancy analysis (dbRDA) showed significant separation between the CD-associated colonic mucosal microbiota and the microbiota of controls (P = 0.035; Fig. 1A). The most abundant genera differences were in Roseburia (decrease), Sutterella (increase), Eubacterium (decrease), and Subdoligranulum (decrease) in CD compared to control (supplementary Table 1, http://links.lww.com/MPG/A116). Interestingly, 9 of the 15 CD samples separated more distantly from controls (Fig. 1A).
Mucosal Granuloma Indicates Distinct Microbiota Separation in CD
When the clinical and histological data of the 9 patients with CD with more distinct microbiota separation were reviewed, the only significant finding was the presence of granuloma or a distinct giant cell in at least 1 biopsy sample from either the upper or the lower gastrointestinal tract in 8 of the 9 patients (Fig. 1A). None of the other 6 patients in the study whose microbiota separated less distinctly had granuloma or giant cell observed. The presence of granuloma significantly predicted a discrete colonic mucosal microbiota separation in CD (P = 0.0014) with 88.8% sensitivity and 100% specificity. dbRDA incorporating the presence or absence of granuloma supported this observation because the microbiota separation was more prominent when this histological parameter was taken into account (P = 0.013, Fig. 1B). In the meantime, principal coordinates analysis (PCoA) did not show significant microbiota divergence between the groups (supplementary Fig 1, http://links.lww.com/MPG/A117) revealing the utility of dbRDA when limited dysbiosis is present in association with disease, or disease subtypes.
Heatmap of the indicator species analysis at the genera (Fig. 2) and the species (supplementary Fig 2, http://links.lww.com/MPG/A117) level showed that patients with CD with granuloma had a higher number of genera and species significantly differentiating the colonic mucosal microbiota from controls than patients without granuloma. Analysis of variance (ANOVA) results for the genera and species comparisons between granuloma-positive, granuloma-negative CD, and controls are presented in supplementary Tables 2 and 3 (http://links.lww.com/MPG/A118).
Although there were 32 genera differing (with Student t test P < 0.05) between granuloma-positive CD patients and controls (supplementary Table 4, http://links.lww.com/MPG/A119), only 4 could be detected in granuloma-negative cases (supplementary Table 5, http://links.lww.com/MPG/A119). None of the 4 genera separating granuloma-negative CD from controls represented >0.4% of the bacterial populations in any of the samples studied.
The most prominent genera distinguishing granulomatous CD from nongranulomatous were Ruminococcus, Roseburia, Eggerthella (decrease), and Porphyromonas (increase) (Fig. 2 and supplementary Table 6 [http://links.lww.com/MPG/A119]). There was a trend for the genera Faecalibacterium to be decreased in the transverse colonic mucosa of granulomatous patients with CD compared with granuloma-negative ones (supplementary Table 6 [http://links.lww.com/MPG/A119]; P = 0.073). Additionally, Dialister (Fig. 2) was absent in 7 of 8 granuloma-positive cases, but detectable in 6 of 7 granuloma-negative CD cases (OR 0.0238 [0.0012–0.4679]; P = 0.0087). On the contrary, Porphyromonas was present in 5 of 8 granulomatous patients with CD, but was undetectable in the granuloma-negative cases (P = 0.026). Histological severity of inflammation in the transverse colon (in adjacent biopsies to ones analyzed for microbiota) was not significantly different between granuloma-positive and -negative cases (P = 0.212). Only the abundance of Ruminococcus showed significant correlation (inverse correlation: supplementary Fig 3, http://links.lww.com/MPG/A117) with microscopic inflammation among the above 6 genera (the others had no significant correlation with inflammation, not shown).
There were no species differentiating granuloma-positive from -negative CD; however, there were 16 species differing between CD and controls (supplementary Table 7, http://links.lww.com/MPG/A119). When the presence of granuloma was taken into account, there were 61 species significantly differing in average abundance between granulomatous CD and controls (supplementary Table 8, http://links.lww.com/MPG/A119) whereas only 18 such could be detected between granuloma-negative CD and controls (supplementary Table 9, http://links.lww.com/MPG/A119). Notably, Ruminococcus gnavus was significantly increased in granuloma-negative CD only (supplementary Table 7, http://links.lww.com/MPG/A119). Faecalibacterium prausnitzii, while detected and distinguished by our microbiota analysis, was not among the significantly partitioning species in any of the comparisons.
Bacterial richness and diversity analyses did not reveal significant differences between control, granuloma-positive, and granuloma-negative groups (supplementary Figs 4–6, http://links.lww.com/MPG/A117).
Fungal Malassezia Genus Associates With Granuloma-positive CD
Fungal dbRDA showed no significant separation between controls and granuloma-positive or -negative cases of CD (supplementary Fig 7, http://links.lww.com/MPG/A117); however, the genus Malassezia was significantly associated with granulomatous CD. ANOVA for fungal genera showed only Malassezia to differentiate significantly between the observed groups (false discovery rate 0.02). A significantly higher proportion of granulomatous patients with CD had Malassezia exceeding a cutoff value of 1% (6 of 7 [1 sample could not be amplified]; OR 25.2 [2.45–259.24], P = 0.0025) compared with control (5 of 26). Samples from granuloma-positive patients with CD more commonly had Malassezia >1% compared to granuloma-negative cases (2 of 7) in the transverse colonic mucosa as well (OR 15 [1.03–218.31], P = 0.102).
MAP Not Found in the Biopsy Specimens
We tested for the presence of MAP by real-time PCR. None of the samples studied was positive.
Significant Geographical Bias
Center and geographical bias was not addressed during the initial evaluations with the intention to identify common microbiota associations (independently from geographical location) of CD in industrialized countries; however, control and disease samples were skewed by sites (Table 1). Therefore, geographical effects on microbiota composition (28) may have significantly influenced our results. Indeed, although PCoA did not separate the sites, dbRDA did (supplementary Fig 8, http://links.lww.com/MPG/A117), indicating potential center bias. Consequently, we separately analyzed the results from the center from which most samples were obtained (Baylor). Geographical/center bias on microbiota composition could have resulted from multiple factors including genetic and dietary differences between the populations studied, variation in bowel-cleansing regimens (29), and differing methods of sample freezing (dry ice vs liquid nitrogen).
Center bias modified the results of bacterial genus and species comparisons between the control, CD, and granuloma-based comparisons. Nevertheless, granuloma-based microbiota separation persisted (P = 0.04; supplementary Fig 9, http://links.lww.com/MPG/A117) and several bacterial genera and species remained different between the groups of the Baylor cohort (supplementary Tables 10–16, http://links.lww.com/MPG/A119). Roseburia and Eubacterium were consistently decreased in CD mucosa (supplementary Table 10, http://links.lww.com/MPG/A119), whereas Roseburia and Ruminococcus (including R bromii, callidus, gnavus, obeum, and as-yet unclassified Ruminococcus species) were less abundant in granulomatous CD than in nongranulomatous, with Porphyromonas being more abundant in the colonic mucosa of granuloma-positive patients (supplementary Table 12, http://links.lww.com/MPG/A119). At the species level, Eubacterium ramulus and Roseburia sp (as-yet unclassified Roseburia sp) were consistently decreased in CD (supplementary Table 13, http://links.lww.com/MPG/A119). Granulomatous CD cases were mostly the reason for this separation with respect to Roseburia sp (supplementary Tables 14 and 16, http://links.lww.com/MPG/A119). Importantly, Dialister was <0.045% abundant in all granulomatous CD samples, whereas >0.07% in all nongranulomatous CD cases in the Baylor cohort (P = 0.0013). Dialister invisus specifically was less abundant (<0.044%) in all granulomatous CD samples, whereas it was present at 0.044% or more in nongranulomatous CD (P = 0.0152).
When the Baylor cohort was analyzed by fungal metagenomics, the association of Malassezia with granulomatous CD was less prominent (5 of 6 CD+ cases >1%, whereas 3 of 9 in control; P = 0.12); however, Saccharomyces (2 of 11 in CD, 8 of 9 in control; P = 0.0055) and Candida (3 of 11 CD, 7 of 9 control; P = 0.07) were surprisingly less frequently detectable in all the Baylor CD samples (same between granulomatous and nongranulomatous) than controls. These findings pertained to Saccharomyces cerevisiae (detectable in 2 of 11 in CD, 8 of 9 in control; P = 0.0055) and Candida albicans (detectable in 1 of 11 in CD, 5 of 9 in control; P < 0.05) as well. These findings warrant further investigation because opposing results (ie, increased abundance of C albicans specifically) were obtained from mouth swabs and stool specimens from patients and family members with CD (30).
Mucosal Granuloma Predicted by CRP Levels in Treatment-naïve Pediatric CD
When the clinical and laboratory parameters of the patients with CD were compared, the patients with granulomatous CD had a trend for higher levels of serum CRP than patients without granuloma. We decided to use an arbitrary cutoff value of CRP (>1 mg/dL). By this means, a CRP >1 mg/dL predicted granulomatous CD by 80% sensitivity and 100% specificity in the microbiota study cohort (P = 0.007). This level of CRP would have forecasted the observed more significant microbiota separation as well (OR 16 [1.09–243.26]; P = 0.089; 80% sensitivity, 80% specificity). The analysis was extended to a larger population of treatment-naïve pediatric CD cases (86 patients, 56% of whom had granuloma/giant cell detected in at least 1 biopsy sample from the upper or the lower intestine). Within this larger population CRP >1 mg/dL significantly predicted granulomatous CD (OR 28 [6–134.32]; P < 0.00001; 73% sensitivity, 91% specificity). The average level of serum CRP was also significantly higher in the granulomatous CD group than in patients without granuloma (P = 0.0059; Fig. 3). CRP testing for the 86 patients was performed in >4 different laboratories; however, CRP for 53 patients was tested in the same laboratory. If we examined only these cases, sensitivity (76%) and specificity (100%) for CRP >1 mg/dL predicting granulomatous CD increased (P < 0.00001) compared to the collective cohort tested in multiple laboratories.
The commensal microbiota is recognized to play an important role in a number of common human disorders including IBD (31–33). In the meantime, it is extremely difficult to overcome or incorporate the tremendous number of confounding variables, which characterize IBD-focused clinical microbiota research (31). Therefore, the pathogenic role of IBD-related dysbiosis and its potential therapeutic implications remain questionable (32).
In the present study we analyzed pediatric transverse colonic mucosal biopsy samples from treatment-naïve CD cases and controls, which enabled us to overcome several of the potential confounding issues of microbiota analysis in IBD, such as examining a clinical subpopulation of patients (9) (early onset, with limited time for chronic inflammation); elimination of treatment bias (19); studying the mucosa-associated microbiota, relevant for intestinal immunomodulation (17); and avoiding colonic location–dependent microbiota variation (16). Several limitations of our analysis are, however, still present: small sample size arising from the strict selection of patients; the nature of the control population, for whom a colonoscopy was indicated and a number of whom may have dysbiosis compared to healthy children regardless of having no obvious intestinal inflammation (34); lack of data regarding specific host genetic factors, such as polymorphisms in IBD susceptibility genes, which can associate with microbiota variation (13,15); and significant center bias according to dbRDA. Nevertheless, our data suggest a microbiota separation within pediatric CD colonic mucosa (Fig. 1) that is modulated by the presence of granulomatous inflammation in any part of the intestinal tract independently from geographic bias (Fig. 1 and supplementary Figure 8, http://links.lww.com/MPG/A117). In a similar study of colonic mucosal samples from adult patients, only about 30% of CD cases separated from controls (35); however, the latter study tested mucosal samples from surgical cases (likely manifesting in patients with long-lasting and/or more aggressive inflammation) without accounting for the effects of treatment, and using a control population of largely (62%) colon cancer patients who themselves may have significant dysbiosis (36). In addition, although we directly tested for differences in the microbiota (using dbRDA), they did not (using PCoA to summarize overall variation). In fact, direct comparison between the PCoAs of the study by Frank et al (35) and ours (supplementary Figure 1, http://links.lww.com/MPG/A117) indicates an even less prominent dysbiosis in treatment-naïve pediatric CD than in the chronically treated, adult, surgical cases. The limited dysbiosis is further emphasized by the fact that none of the direct taxa comparisons in the present study were significant after correcting for multiple tests (supplementary Tables 1–16, http://links.lww.com/MPG/A116, http://links.lww.com/MPG/A118, http://links.lww.com/MPG/A119); however, this latter result may be a consequence of the small sample sizes originating from the challenges in obtaining treatment-naïve samples. Therefore, the direct taxa comparisons represent only trends, which need to be verified in further studies with larger sample sizes.
We also detected a significant correlation between granulomatous CD and colonic mucosal microbiota variation by dbRDAs. Granuloma is a pathognomic feature of CD in the clinical setting of IBD and can be detected in 21% to 60% of patients, with a higher frequency in children (10,37,38). The pathogenesis of granuloma formation in CD is unknown, although some studies hypothesize that specific bacterial components may play a role (39). Granulomas in pediatric CD have been associated with an increased incidence of perianal disease and gastritis (10). MAP has been detected in a higher number of pediatric patients with granulomatous CD than in controls (40), but these observations were not confirmed with similar methodology from adult patients with CD in biopsies containing mucosal granuloma (12). Therefore, the importance of MAP in regards to IBD pathogenesis remains questionable.
Roseburia was decreased in the colonic mucosa of our CD proband, which was largely attributable to the subjects with granuloma irrespective of geographic bias. This was not observed by Willing et al (14), who only found significant decreases of this genus in the ileum and feces of patients with ileal CD, but not in the mucosa of the large bowel of patients with colonic CD, perhaps indicating that patients with granulomatous CD were more common in the ileal CD group of their cohort than in the group with colonic disease. Interestingly, a decrease in Roseburia was recently found in the colonic mucosa of young adult mice sensitive to experimental colitis secondary to maternal dietary modification (41), and in patients with UC as well (42).
A decrease in Ruminococcus (supplementary Tables 4, 6, 12, http://links.lww.com/MPG/A119) was detected in our patients with granulomatous CD, which was also correlated with the severity of mucosal inflammation (supplementary Fig 6, http://links.lww.com/MPG/A117). Similar results were obtained from the colonic mucosa of adults (43), but not controlled for therapy, granulomas, or inflammation. Perhaps this may be the reason for the discordance between our results and this later manuscript in respect to Eubacterium, where we found a CD-associated decrease (supplementary Table 3, http://links.lww.com/MPG/A118) as opposed to Verma et al (43).
Although we could associate granulomatous CD with the decreased presence of Dialister, and increased detection of Porphyromonas and Malassezia, none of these bacterial and fungal taxa were exclusive for differentiating this CD phenotype when examined in all the samples. Malassezia can induce granulomatous inflammation (44), and at least 1 case has been reported of Malassezia furfur sepsis in a patient with CD (45); however, when geographical bias was considered, the association of Malassezia with granulomatous CD was less prominent. In the meantime, Dialister separated all granulomatous samples with being less abundant than in nongranulomatous CD from the Baylor cohort, which pertained to D invisus specifically as well. A decrease in D invisus has been observed in stool from patients with CD and their relatives, supporting our findings (46).
Our results emphasize that giant cells and granulomas in CD are important associates of microbiota composition and should be incorporated into future work in the metagenomics of this disorder.
The finding of CRP levels >1 mg/dL correlating with granulomatous CD and microbiota separation in treatment-naïve patients is also novel and important for indicating this molecule as a potential biomarker for host-microbial interactions in this disease group. Previous investigations did not reveal association between granulomatous CD and CRP (10), but it is unclear whether this was selectively examined in treatment-naïve patients, and a cutoff point of significance was not established.
The present study includes the first high-throughput microbiota analysis of treatment-naïve CD incorporating bacterial and fungal metagenomics on colonic mucosal specimens. Therefore, it can be considered as pilot study for future, larger-scale, high-throughput metagenomic investigations in treatment-naïve intestinal samples from patients with IBD. Our findings support that stringent, geographic/center, clinical, molecular, and histological selection of patients can further our understanding of dysbiosis in IBD and the relation between the microbiota and immune responses.
The authors thank the patients and their families for participating in the study, and Antone R. Opekun for assistance.
1. Loftus EV Jr. Clinical epidemiology of inflammatory bowel disease: incidence, prevalence, and environmental influences. Gastroenterology 2004; 126:1504–1517.
2. Molodecky NA, Soon IS, Rabi DM, et al. Increasing incidence and prevalence of the inflammatory bowel diseases with time, based on systematic review. Gastroenterology 2012; 142:46e42–54e42.
3. Braegger CP, Ballabeni P, Rogler D, et al. Epidemiology of inflammatory bowel disease: is there a shift towards onset at a younger age? J Pediatr Gastroenterol Nutr 2011; 53:141–144.
4. Lewis JD. The utility of biomarkers in the diagnosis and therapy of inflammatory bowel disease. Gastroenterology 2011; 140:1817e2–1826e2.
5. Solem CA, Loftus EV Jr, Tremaine WJ, et al. Correlation of C-reactive protein with clinical, endoscopic, histologic, and radiographic activity in inflammatory bowel disease. Inflamm Bowel Dis 2005; 11:707–712.
6. Henriksen M, Jahnsen J, Lygren I, et al. C-reactive protein: a predictive factor and marker of inflammation in inflammatory bowel disease. Results from a prospective population-based study. Gut 2008; 57:1518–1523.
7. Jurgens M, Mahachie John JM, Cleynen I, et al. Levels of C-reactive protein are associated with response to infliximab therapy in patients with Crohn's disease. Clin Gastroenterol Hepatol 2011; 9:421e1–427e1.
8. Kelsen J, Baldassano RN. Inflammatory bowel disease: the difference between children and adults. Inflamm Bowel Dis 2008; 14 (suppl 2):S9–S11.
9. Van Limbergen J, Russell RK, Drummond HE, et al. Definition of phenotypic characteristics of childhood-onset inflammatory bowel disease. Gastroenterology 2008; 135:1114–1122.
10. De Matos V, Russo PA, Cohen AB, et al. Frequency and clinical correlations of granulomas in children with Crohn disease. J Pediatr Gastroenterol Nutr 2008; 46:392–398.
11. Ryan P, Kelly RG, Lee G, et al. Bacterial DNA within granulomas of patients with Crohn's disease—detection by laser capture microdissection and PCR. Am J Gastroenterol 2004; 99:1539–1543.
12. Toracchio S, El-Zimaity HM, Urmacher C, et al. Mycobacterium avium subspecies paratuberculosis and Crohn's disease granulomas. Scand J Gastroenterol 2008; 43:1108–1111.
13. Frank DN, Robertson CE, Hamm CM, et al. Disease phenotype and genotype are associated with shifts in intestinal-associated microbiota in inflammatory bowel diseases. Inflamm Bowel Dis 2011; 17:179–184.
14. Willing BP, Dicksved J, Halfvarson J, et al. A pyrosequencing study in twins shows that gastrointestinal microbial profiles vary with inflammatory bowel disease phenotypes. Gastroenterology 2010; 139:1844e1–1854e1.
15. Rehman A, Sina C, Gavrilova O, et al. Nod2 is essential for temporal development of intestinal microbial communities. Gut 2011; 60:1354–1362.
16. Eckburg PB, Bik EM, Bernstein CN, et al. Diversity of the human intestinal microbial flora. Science 2005; 308:1635–1638.
17. Swidsinski A, Ladhoff A, Pernthaler A, et al. Mucosal flora in inflammatory bowel disease. Gastroenterology 2002; 122:44–54.
18. Nagalingam NA, Lynch SV. Role of the microbiota in inflammatory bowel diseases. Inflamm Bowel Dis 2012; 18:968–984.
19. Lepage P, Hasler R, Spehlmann ME, et al. Twin study indicates loss of interaction between microbiota and mucosa of patients with ulcerative colitis. Gastroenterology 2011; 141:227–236.
20. Conte MP, Schippa S, Zamboni I, et al. Gut-associated bacterial microbiota in paediatric patients with inflammatory bowel disease. Gut 2006; 55:1760–1767.
21. Poulain D, Sendid B, Standaert-Vitse A, et al. Yeasts: neglected pathogens. Dig Dis 2009; 27 (suppl 1):104–110.
22. Momozawa Y, Deffontaine V, Louis E, et al. Characterization of bacteria in biopsies of colon and stools by high throughput sequencing of the V2 region of bacterial 16S rRNA gene in human. PLoS One 2011; 6:e16952.
23. Satsangi J, Silverberg MS, Vermeire S, et al. The Montreal classification of inflammatory bowel disease: controversies, consensus, and implications. Gut 2006; 55:749–753.
24. Bailey MT, Walton JC, Dowd SE, et al. Photoperiod modulates gut bacteria composition in male Siberian hamsters (Phodopus sungorus). Brain Behav Immun 2010; 24:577–584.
25. Dowd SE, Zaragoza J, Rodriguez JR, et al. Windows.NET network distributed basic local alignment search toolkit (W.ND-BLAST). BMC Bioinformatics 2005; 6:93.
26. Mills DK, Entry JA, Voss JD, et al. An assessment of the hypervariable domains of the 16S rRNA genes for their value in determining microbial community diversity: the paradox of traditional ecological indices. FEMS Microbiol Ecol 2006; 57:496–503.
27. Dowd SE, Delton Hanson J, Rees E, et al. Survey of fungi and yeast in polymicrobial infections in chronic wounds. J Wound Care 2011; 20:40–47.
28. Fallani M, Young D, Scott J, et al. Intestinal microbiota of 6-week-old infants across Europe: geographic influence beyond delivery mode, breast-feeding, and antibiotics. J Pediatr Gastroenterol Nutr 2010; 51:77–84.
29. Harrell L, Wang Y, Antonopoulos D, et al. Standard colonic lavage alters the natural state of mucosal-associated microbiota in the human colon. PLoS One 2012; 7:e32545.
30. Standaert-Vitse A, Sendid B, Joossens M, et al. Candida albicans colonization and ASCA in familial Crohn's disease. Am J Gastroenterol 2009; 104:1745–1753.
31. Young VB, Kahn SA, Schmidt TM, et al. Studying the enteric microbiome in inflammatory bowel diseases: getting through the growing pains and moving forward. Front Microbiol 2011; 2:144.
32. Sartor RB. Key questions to guide a better understanding of host-commensal microbiota interactions in intestinal inflammation. Mucosal Immunol 2011; 4:127–132.
33. Nagy-Szakal D, Kellermayer R. The remarkable capacity for gut microbial and host interactions. Gut Microbes 2011; 2:178–182.
34. Saulnier DM, Riehle K, Mistretta TA, et al. Gastrointestinal microbiome signatures of pediatric patients with irritable bowel syndrome. Gastroenterology 2011; 141:1782–1791.
35. Frank DN, St Amand AL, Feldman RA, et al. Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases. Proc Natl Acad Sci U S A 2007; 104:13780–13785.
36. Sobhani I, Tap J, Roudot-Thoraval F, et al. Microbial dysbiosis in colorectal cancer (CRC) patients. PLoS One 2011; 6:e16393.
37. Molnar T, Tiszlavicz L, Gyulai C, et al. Clinical significance of granuloma in Crohn's disease. World J Gastroenterol 2005; 11:3118–3121.
38. Rubio CA, Orrego A, Nesi G, et al. Frequency of epithelioid granulomas in colonoscopic biopsy specimens from paediatric and adult patients with Crohn's colitis. J Clin Pathol 2007; 60:1268–1272.
39. Pierik M, De Hertogh G, Vermeire S, et al. Epithelioid granulomas, pattern recognition receptors, and phenotypes of Crohn's disease. Gut 2005; 54:223–227.
40. Lee A, Griffiths TA, Parab RS, et al. Association of Mycobacterium avium subspecies paratuberculosis with Crohn disease in pediatric patients. J Pediatr Gastroenterol Nutr 2011; 52:170–174.
41. Schaible TD, Harris RA, Dowd SE, et al. Maternal methyl-donor supplementation induces prolonged murine offspring colitis susceptibility in association with mucosal epigenetic and microbiomic changes. Hum Mol Genet 2011; 20:1687–1696.
42. Vermeiren J, Van den Abbeele P, Laukens D, et al. Decreased colonization of fecal Clostridium coccoides/Eubacterium rectale species from ulcerative colitis patients in an in vitro dynamic gut model with mucin environment. FEMS Microbiol Ecol 2012; 79:685–696.
43. Verma R, Verma AK, Ahuja V, et al. Real-time analysis of mucosal flora in patients with inflammatory bowel disease in India. J Clin Microbiol 2010; 48:4279–4282.
44. Desai HB, Perkins PL, Procop GW. Granulomatous dermatitis due to Malassezia sympodialis. Arch Pathol Lab Med 2011; 135:1085–1087.
45. Bhargava P, Longhi LP. Images in clinical medicine. Peripheral smear with Malassezia furfur. N Engl J Med 2007; 356:e25.
46. Joossens M, Huys G, Cnockaert M, et al. Dysbiosis of the faecal microbiota in patients with Crohn's disease and their unaffected relatives. Gut 2011; 60:631–637.