O-025 Paneth Cell Phenotype Correlates with Genetics, Transcriptome Profile, Pathologic Hallmark and Predicts Clinical Outcome in Patients with Crohnʼs Disease

Abstract

The identification of genetic susceptibility loci in Crohn's Disease (CD) represents an opportunity to define disease subtypes that could be utilized in the development of novel personalized therapies. The major challenges are that the CD susceptibility loci are numerous, complex and likely interact with undefined components of the environment. Paneth cells in the small intestine form part of the innate immunity and play crucial roles in CD pathogenesis. We have previously shown that the morphology of Paneth cell cytoplasmic antimicrobial granules can be used as a readout for host genetics and environmental exposure. We hypothesized that Paneth cell phenotypes could act as an integrative readout to define specific CD subtypes. We retrospectively analyzed 178 genotyped CD patients with small intestinal resections at 2 institutions. Uninvolved ileal sections were analyzed for transcriptional profiles by microarray and Paneth cell phenotype by lysozyme immunofluorescence (up to 250 crypts quantified/sample). The morphology of each individual Paneth cell was categorized. The sum of Paneth cells with abnormal lysozyme distribution was used to determine the patients' phenotype, with “bad” Paneth cell phenotype defined as ≥20% abnormal Paneth cells, and <20% defined as “good” Paneth cell phenotype. The patients' phenotypes were then correlated with genotype, pathologic features, transcriptional profiles, and time to recurrence after surgery. Among the 178 CD patients, 27 (15%) had bad Paneth cell phenotype. Patients with one or more CD-associated NOD2 susceptibility alleles had increased proportions of abnormal Paneth cells compared to those without NOD2 susceptibility alleles. Furthermore, there is an additive effect of CD-associated NOD2 susceptibility alleles and ATG16L1 T300A on Paneth cell abnormalities. We found that the cases with bad Paneth cell phenotype were more likely to be granuloma-poor (P = 0.0160). Bad Paneth cell phenotype was also associated with an altered transcriptional signature most strongly associated with immune activation pathways. Clinically, among the 121 patients who received adjuvant prophylactic therapy, those with bad Paneth cell phenotype showed shorter time to disease recurrence after surgery (median 10 months) compared to patients with good Paneth cell phenotype (median 14 months; P = 0.0198). The difference was not seen in patients who did not receive adjuvant prophylactic therapy. We also found that in the same patients, the Paneth cell phenotype in involved area was consistent with that in the uninvolved area (P = 0.0001). Finally, bioequivalence analysis showed that a minimum of 40 crypts in biopsy was sufficient for Paneth cell morphology analysis. Paneth cell phenotype is associated with patients' genetics, pathologic hallmark, and predicts clinical outcome. It can be used as a “hard” endpoint that can subdivide CD patients into clinically relevant groups. Defining the minimum crypt number required for Paneth cell phenotype analysis enables the use of routine biopsy material for further clinical studies.