A181 HIGHER CUMULATIVE HISTOLOGIC INFLAMMATORY BURDEN SCORE IS ASSOCIATED WITH THE RISK OF DEVELOPMENT OF COLORECTAL NEOPLASIA IN ULCERATIVE COLITIS

Background

Ulcerative colitis (UC) patients have an elevated risk of colorectal neoplasia (CRN). Younger age at diagnosis, extent of colitis, and longer duration of colitis, as well as increased severity, which is a component of the cumulative inflammatory burden score (CIB), have been associated with the development of CRN. CIB was developed based on a large cohort of UC patients from St. Mark’s Hospital (UK) but needs further validation in independent cohorts.

Purpose

We analyzed the association between higher histologic CIB and development of CRN in longstanding UC patients.

Method

A matched case-control cohort of UC patients with at least 8 years of disease duration was analyzed at Mount Sinai Hospital. Patients with primary sclerosing cholangitis were excluded. Cases consisted of UC patients with colitis-associated neoplasia defined as indefinite for dysplasia (IND), low-grade dysplasia (LGD), high-grade dysplasia (HGD), or colorectal cancer (CRC). Each case was matched to two controls by age at disease onset, disease duration, and histological extent of colitis. Histologic reports obtained by colonoscopy were reviewed and histological activity was assessed as quiescent/normal (0), mild (1), moderate (2), and severe (3). The colonic area with the higher score was utilized and the CIB was calculated by summing each score and multiplying it by the interval of surveillance. A mean CIB (mCIB) was also calculated for each patient dividing the CIB by the number of colonoscopies. Continuous variables including CIB scores and mCIB scores were summarized as median and interquartile range (IQR) and differences between groups were compared by Mann-Whitney test.

Result(s)

Fifty-four UC patients were analyzed with 18 having CRN (6 CRC, 2 HGD, 3 LGD and 7 IND) and 36 controls without CRN. The clinical characteristics of the total cohort, cases and controls are depicted in the Table. Median age at last colonoscopy assessed was 45 years (36-55) and 40.7% were female. The median age at onset of UC was 23 years (19-37) and median duration of UC was 16 years (11-23). All patients had extensive histologic colonic disease. There were no differences between cases and controls in interval of surveillance evaluated (7.5 vs 7.8 years, p = 0.7) and median number of colonoscopies with histologic assessment (4 vs 4, p =0.6). Cases with CRN had significantly higher CIB (11.4 vs 7.9, p = 0.02) and mCIB (2.9 vs 2.0, p = 0.02) compared to controls.

Image

Conclusion(s)

The histologic CIB score is associated with an increased risk of developing CRN in UC patients with similar age at onset of disease, disease duration and colitis extent. Given CIB reflects the severity of histologic inflammation over the years, treatment strategies to improve histologic inflammation could reduce the incidence of CRN in UC.

Please acknowledge all funding agencies by checking the applicable boxes below

None

Disclosure of Interest

None Declared

A235 ILEAL MICROBIOME ALPHA DIVERSITY REMAINS LOWER IN ENDOSCOPICALLY AND HISTOLOGICALLY INACTIVE CROHN’S DISEASE COMPARED TO ULCERATIVE COLITIS OR HEALTHY CONTROLS

Background

Crohn’s disease (CD) has been associated with a lower alpha diversity when compared to ulcerative colitis (UC) patients and healthy controls (HC), which also depends on disease location and endoscopic activity. However, it is unclear whether the resolution of histologic inflammation may influence mucosal alpha diversity in the ileum.

Aims

To characterize the ileal mucosa-associated microbiome diversity in subjects with ileal predominant CD (iCD) in endoscopic (ER) and endo-histologic (EHR) remission, compared to subjects with colonic predominant CD (cCD), UC and HC, respectively.

Methods

Data from a large cohort of subjects recruited at Mount Sinai Hospital Toronto (2009–2016) was analyzed. ER was defined as a segmental Simple Endoscopic Score for CD of <3 in the ileum. EHR was defined as ER with the absence of active histologic inflammation in ileal biopsies. CD patients were divided according to the Montreal classification into iCD (L1 and L3) and cCD (L2) and compared against UC patients without backwash ileitis and HC. Patients with history of ileocecal resection and/or antibiotic use at baseline were excluded. Microbial 16S rRNA gene was sequenced using the Illumina MiSeq and processed using QIIME2. Alpha diversity was measured using the Shannon index and compared using Kruskal-Wallis test and further pairwise Wilcoxon with Holm correction. An adjusted p-value < 0.05 was considered significant.

Results

We included 35 CD patients with ileal ER, 81 UC patients and 32 HC. Among CD patients, 20/35 (57.1%) and 15/35 (42.9%) had iCD and cCD, respectively. Ileal mucosal alpha diversity was significantly lower in iCD patients in ER compared with that of UC patients (q=0.004) and HC (q=0.001). No differences in ileal mucosal alpha-diversity were seen between iCD than cCD patients (q=0.12). When histology was included to classify CD and UC patients as EHR, the ileal alpha diversity of iCD patients remained reduced compared to UC patients and HC (q=0.008 and q=0.002, respectively). Again, ileal mucosal alpha-diversity was not significantly lower in iCD than cCD patients among those in ileal EHR (q=0.24).

Conclusions

Ileal mucosa of CD patients in ER and EHR have lower alpha diversity than UC and HC. These findings suggest that the ileal mucosa of CD patients remains dysbiotic despite achieving endoscopic and histologic remission.

Funding Agencies

IBD Genetics Consortium

A238 ALTERED GUT MICROBIOME COMPOSITION AND FUNCTION ARE ASSOCIATED WITH GUT BARRIER DYSFUNCTION IN HEALTHY RELATIVES OF CROHN’S DISEASE PATIENTS

Background

The gut microbiome may play a role in gut barrier homeostasis including epithelial barrier function, but data are scarce and limited to animal studies

Aims

To assess if alterations in gut microbiome are associated with gut barrier function

Methods

We utilized the Genetic Environmental Microbial (CCC GEM) cohort of healthy first-degree relatives (FDRs) of Crohn’s disease (CD) patients. Gut barrier function was assessed using the ratio of urinary fractional excretion of lactulose to mannitol (LMR). Stool bacterial DNA was extracted and sequenced for the V4 hypervariable region of the 16S rRNA gene using MiSeq and processed using QIIME2. Microbial functions were imputed using PICRUSt2. The cohort was divided into a North American discovery cohort (n=2,472) and non-North American external validation cohort (n=655). LMR>0.025 was defined as abnormal. LMR-microbiome associations were assessed using multivariable regression model and Random Forest (RF) classifier algorithm. q<0.05 was considered significant when multiple tests were performed

Results

The median age of the entire cohort was 17.0 years [IQR 12.0; 24.0], 52.6% were females and 25.4% had LMR>0.025. In the discovery cohort, subjects with LMR>0.025 had markedly reduced alpha diversity (Chao1 index, estimate= -0.0037, p=4.0e-04) and altered beta diversity (Bray-Curtis dissimilarity index, PERMANOVA: pseudo-F statistic = 2.99, p=1.0e-03). We identified eight bacterial genera and 52 microbial pathways associated with LMR>0.025 (q<0.05). Four genera (decreased Adlercreutzia [odds ratio(OR)=0.74, 95% confidence interval (CI) 0.6–0.91], Clostridia-UCG-014 [OR=0.71, 95%CI 0.59–0.86], and Clostridium-sensu-stricto-1 [OR=0.75, 95%CI 0.61–0.92] and increased Colidextribacter [OR=1.65, 95%CI 1.2–2.26]) and eight pathways (including decreased biosynthesis of glutamate [OR=0.4, 95%CI 0.21–0.74], tryptophan [OR=0.06, 95%CI 0.01–0.27] and threonine [OR=0.038, 95%CI 0.003–0.41]) were replicated. Bacterial community composition was associated with gut barrier homeostasis as defined by the RF analysis (p= 1.4e-6)

Conclusions

Gut microbiome community and pathways are associated with gut barrier function. These findings may identify potential microbial targets to modulate barrier function

Submitted on behalf of the CCC-GEM Consortium

Funding Agencies

CCC, CIHRCrohn’s and Colitis Canada Genetics Environment Microbial (CCC-GEM) III; The Leona M. and Harry B. Helmsley Charitable Trust; Kenneth Croitoru is the recipient of the Canada Research Chair in Inflammatory Bowel Diseases

A161 MUCOSAL TRANSCRIPTOMICS IN NON-INFLAMED ILEUM OF CROHN’S DISEASE PATIENTS SHOWS DIFFERENTIALLY EXPRESSED GENE PROFILE COMPARED TO ILEUM OF HEALTHY CONTROLS

Background

The terminal ileum is the most susceptible location to develop Crohn’s disease (CD) and therefore is a valuable tissue to investigate biological mechanisms underlying chronic inflammation. Gene expression is highly affected by the level of inflammation, however, data on ileal transcriptomic profile in the absence of active inflammation is limited.

Aims

To investigate the ileal mucosal transcriptomic profile of CD patients in endoscopic and histologic remission compared to healthy controls (HC).

Methods

Ileal biopsies were collected during colonoscopy from patients with CD and HC. Biopsies were classified as non-inflamed based on endoscopic appearance and histologic criteria. Endo-histologic remission (EHR) was defined as simple endoscopic score < 3 and absence of active histologic inflammation. We included CD patients with EHR and HC for the analysis of this study. CD phenotype was divided into isolated colonic CD (cCD) that included Montreal L2 and ileal predominant CD (iCD) that included Montreal L1 and L3. Total RNA was extracted from samples, sequenced using a HiSeq 2500 instrument (Illumina, San Diego, CA, USA) and differential expression analysis was performed in EdgeR. Genes that were differentially expressed at the average of 2-fold-change (FC) in mean expression and False Discovery Rate (FDR) < 0.05 were considered significant.

Results

Ileal samples from 14 CD patients in EHR and 29 HC were included in the analysis. CD patients were significantly younger (median age 28.5 years, interquartile range (IQR)=24–40) compared with HC (median age 56 years, IQR=51–64). There were no differences in gender distribution (42.9% males in CD and 55.2% males in HC). We found 101 differentially expressed genes in CD patients compared to HC (99 genes were up-regulated and two were down-regulated). Dual oxidase 2 (DUOX2) and complement C6 were respectively the most significant up-regulated (logFC=4.45, FDR=1.3e-8) and down-regulated (logFC=-2.73, FDR=0.0006) genes in non-inflamed ileum CD group when compared with HC. In a subgroup analysis comparing ileal samples of CD patients with cCD (n=8) versus iCD phenotype (n=6), no differentially expressed genes were identified.

Conclusions

The mucosal transcriptomic profile of patients with inactive endoscopic and histologic ileal CD shows significant differentially expressed gene profile compared to ileal mucosa of HC. The clinical relevance of these findings should be further investigated.

Funding Agencies

IBD Genetics Consortium

A156 MUCOSA-ASSOCIATED MICROBIOTA OF ILEOCOLONIC CROHN’S DISEASE PATIENTS IS DISTINCT FROM COLONIC CROHN’S DISEASE AND ULCERATIVE COLITIS PATIENTS INDEPENDENT OF BIOPSY SITE, ENDOSCOPIC INFLAMMATION AND HOST GENETICS

Background

Colonic IBD encompassing ulcerative colitis (UC) and isolated colonic Crohn’s disease (cCD) shows significant clinical, therapeutic response and genetic differences compared to ileocolonic CD (icCD). Elucidating the microbial signatures characterizing these subphenotypes could help to understand the causal factors underlying these clinical dissimilarities

Aims

We compared the mucosal microbial diversity and differential abundance (DA) among disease locations (UC, cCD and icCD) accounting for potential clinical, endoscopic, and genetic confounders

Methods

Healthy control (HC), UC, cCD and icCD patients (including ileal and ileocolonic involvement) underwent colonoscopy. Biopsy samples were obtained from terminal ileum (TI), ascending colon (AC) and sigmoid colon (SC) for 16s rRNA gene profiling. Patients with prior ileocecal resection, IBD-unclassified and antibiotic exposure within 3 months before colonoscopy were excluded. Endoscopic inflammation was defined as a segmental Mayo endoscopic subscore = 0 in UC and a simple endoscopic score ≤ 2 in CD. A blood sample was drawn for genotyping and a weighted genetic risk score (GRS) was built based on 169 IBD risk variants found in our cohort. Alpha diversity (Chao1) and DA between IBD subphenotypes were compared using a linear mixed-effects model with subjects as random effect and adjusted for biopsy site, endoscopic inflammation, age, sex, and GRS. For DA analysis, the MaAsLin2 protocol was applied. All p-values were corrected by false discovery rate (FDR) with < 0.05 considered significant

Results

A total of 199 IBD patients and 44 HC with a mean age of 37.2 ± 14 were recruited. Of these, 113 (46.5%) were female. At colonoscopy, 535 biopsy samples (TI = 178, AC = 123 and SC = 234) were obtained. Considering disease location, 254, 55 and 148 samples were obtained from UC, cCD and icCD patients, respectively. A total of 168 samples (31.4%) showed endoscopic inflammation. Alpha diversity was significantly reduced in icCD when compared to either HC, UC or cCD. MaAsLin2 identified that the genera Agathobacter and Faecalibacterium, as well as the family Ruminococacceae and the order Oscillospirales were significantly reduced in icCD when compared to either HC, UC or cCD. These findings were independent of age, sex, endoscopic inflammation, biopsied site, and GRS. UC and cCD did not show differences in their microbial profile

Conclusions

Mucosal samples from UC and cCD patients showed marked similarities in their microbial profile while icCD is characterized by a significant decrease in diversity and beneficial microbes. These data suggest that disease location is the main driver of the mucosal microbial landscape independent of IBD GRS

Funding Agencies

NoneNIDDK IBD Genetics Consortium

A9 BACTERIAL BILE SALT HYDROLASE GENE ABUNDANCE IS ASSOCIATED WITH RORC GENE EXPRESSION IN INTESTINAL MUCOSA OF INFLAMMATORY DISEASE PATIENTS

Background

The role of gut microbes involved in bile acid metabolism and their impact on mucosal immune regulation is beginning to be appreciated. For instance, changes in microbial bile salt hydrolase (BSH) activity which deconjugates bile acids in the gastrointestinal tract of gnotobiotic mice, significantly alters gene expression patterns of immune-related genes in ileum. Moreover, bile acid dysmetabolism may participate in the chronic inflammation loop of Inflammatory bowel disease (IBD).

Aims

We carried out an integrated mucosal microbiome-transcriptome analysis to elucidate associations between microbial bile-acid metabolizing function and host gene expression.

Methods

Crohn’s disease (CD), ulcerative colitis (UC) and IBD unclassified (IBDU) patients were recruited prior to scheduled colonoscopy performed as part of clinical care. Only patients with non-inflamed mucosa defined as a segmental simple endoscopic score 0–2 in CD and a segmental Mayo endoscopic score of 0 in UC/IBDU were included in this analysis to minimize the effect of inflammation on gene expression. Biopsy samples were obtained from terminal ileum, ascending colon and sigmoid colon, and microbial DNA and human RNA was extracted. V4 region of 16S rRNA gene was sequenced and the relative abundance of bile acid-metabolizing genes was inferred using PICRUSt. RNA-seq was used to sequence total human RNA and a supervised transcript reduction analysis focus upon 65 genes previously associated with bile acid metabolism and IBD was utilized. Associations between microbiome clusters of orthologous groups (COGs), transcriptome, diagnosis (CD vs UC/IBDU), and biopsy site were analyzed using linear mixed-effects model with lmer4 function in R. An adjusted-p value after false discovery rate correction &lt; 0.05 was considered significant.

Results

A total of 126 samples from 86 subjects were analyzed corresponding to 35 CD and 51 UC/IBDU. Mean age for the total cohort was 34.7 ± 11 years and 35 (40.6%) were females. There was a significant negative correlation between relative abundance of bacterial bsh genes (COG3049) and human RORC gene (p &lt; 0.03). This association was independent of type of diagnosis and biopsy site. There was no association among other analyzed bacterial COGs and host genes.

Conclusions

Using an integrative microbiome-host transcriptome approach, our data provide new evidence linking microbial bile acid deconjugation (bsh genes) and host gene expression in the mucosal-luminal interface in quiescent IBD-affected tissue. Nuclear receptor RORC is pivotal in the differentiation and function of innate lymphoid cells and T-helper 17 cells. Modulation of this pathway by bile acids or gut bacteria involved in their metabolism could shed light on the immune role of bile acids in IBD patients.

Funding Agencies

CAG, CIHRNational Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)

A41 LOSS OF SATB2 EXPRESSION IN COLORECTAL CANCER IS ASSOCIATED WITH DURATION OF INFLAMMATORY BOWEL DISEASE

Background

Patients with inflammatory bowel disease (IBD) are at higher risk for developing colitis-associated colorectal cancer (CAC). Clinical and endoscopic features are used to stratify the risk of CAC, but new biomarkers are necessary to improve this stratification. Recent studies have shown that loss of expression of special AT-rich sequence binding protein 2 (SATB2) is frequent in CAC compared to sporadic colorectal cancer and this SATB2 status is found in pre-cancerous dysplastic lesions as well. However, the relationship of known clinical risk factors for CAC and loss of SATB2 has not been explored.

Aims

To assess the association of loss of SATB2 expression in CAC with clinical characteristics of IBD.

Methods

Patients with a known diagnosis of ileocolonic or colonic Crohn’s disease (CD), ulcerative colitis (UC), or IBD unclassified (IBDU) who underwent colectomy between October 2010 and December 2017 for CAC were included. SATB2 expression in neoplastic tissue was evaluated using immunohistochemistry (IHC), where less than 5% of tumor cells showing staining was considered loss of SATB2. Tumor grade, P53 and mismatch repair (MMR) status were assessed as well. Available clinical data such as sex, smoking status, IBD diagnosis (CD, UC or IBDU), age at IBD diagnosis, duration of IBD, extent of colitis and previous medications were collected. We used a generalized linear model to assess the association between these biomarkers and clinical data.

Results

A total of 58 patients with mean age at CAC diagnosis of 50.3 ±13 years, 27 (46%) females were analyzed. Mean IBD duration was 17.6 ±10 years and 22 (37.9%), 34 (58.6%) and 2 (3.4%) were CD, UC and IBDU, respectively. Thirty-two (55.2%) CACs had loss of SATB2 expression. There was no association between age at CAC diagnosis or grade of the tumor and loss of SATB2. However, longer duration of IBD (21.2 ± 9 years vs 13.7 ± 9 years, p = 0.01) was significantly associated with loss of SATB2. There was no association between SATB2 status and other explored clinical or endoscopic variables. Tumors with P53 mutation were associated with a younger age at diagnosis of CAC (47.2 ±13 vs 55.0 ±12 years, p = 0.03), but no other associations of this marker or MMR with clinical or endoscopic variables of IBD were found.

Conclusions

Loss of SATB2 expression is significantly associated with IBD duration, a well-known risk factor for CAC. This association with duration of IBD could denote an effect of longer chronic inflammation on SATB2 status. Given the previously reported association of loss of expression of SATB2 with pre-cancerous lesions in IBD patients, this could be a potential biomarker for risk of CAC.

Funding Agencies

National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)