To identify and develop novel targeted therapies for complex diseases such as Crohn's disease (CD), functional subtypes rooted in pathogenesis must be defined. One candidate method to subtype CD is to define the small intestinal Paneth cell phenotypes based on the intracellular distribution of antimicrobial proteins. We have previously shown that in CD patients, Paneth cell phenotype correlates with the patients' genotype, distinct gene expression signature, presence of granuloma (pathologic hallmark), and time to recurrence after surgery. However, the mechanism by which abnormal Paneth cells contribute to pathogenesis and sub-classify disease in the context of host-microbial interaction is unclear. We analyzed Paneth cell phenotype and its correlation with mucosal microbiome and transcriptome in a cohort of pediatric CD and non-inflammatory bowel disease (IBD) patients.

We first retrospectively analyzed Paneth cell phenotypes using archived resection specimens from adult (n = 531) and pediatric (n = 73) CD patients. We next analyzed a prospectively recruited pediatric cohort, including CD (n = 44) and non-IBD (n = 62) patients aged 4 to 18. These patients were recruited at the time of routine endoscopy. Ileal mucosal biopsy samples were collected and analyzed for Paneth cell phenotype (lysozyme/defensin 5 immunofluorescence), mucosal microbiome (16S rRNA sequencing), and transcriptome (RNA-sequencing). Paneth cell phenotype was determined by the percentage of normal Paneth cells in each sample. Type I Paneth cell phenotype was defined as <80% normal Paneth cells, whereas Type II Paneth cell phenotype was defined as ≥80% normal Paneth cells.

The prevalence of type I Paneth cell phenotype in pediatric CD patients was higher than in adult CD cohorts (47% versus 18%; P < 0.0001). In pediatric CD patients, the type I Paneth cell phenotype was associated with significant changes in the ileal mucosal microbiome, characterized by reduced abundance of barrier-associated microbes (Faecalibacterium, Blautia, Ruminococcaceae, Porphyromonas, Lachnospira, Peptostreptococcus, Anaerostipes, and Odoribacteraceae) and enrichment of potentially pro-inflammatory microbes (Corynebacterium and Erysipelotrichaceae). In addition, pediatric CD patients with type I Paneth cell phenotype also displayed an altered epithelial gene expression profile, with significant reduction in oxidative phosphorylation gene cluster. Furthermore, the down regulation of oxidative phosphorylation gene cluster in CD was in turn associated with reduced abundance of Faecalibacterium, suggesting a complex network between Paneth cell function, epithelial energy/metabolism, and microbiome homeostasis. The connections between Paneth cell phenotypes, microbiome, and transcriptome profiles were not observed in non-IBD patients.

These data support a functional role for Paneth cells in subtyping CD that is based on a specific pattern of metabolic dysregulation and mucosal dysbiosis.