A five-node causal circuit for ileal Crohn's disease

This updated preprint presents a structured and falsifiable causal framework for ileal-predominant Crohn’s disease. Although more than 200 genetic loci have been associated with Crohn’s, effective therapies continue to converge on a limited number of immune pathways. This discrepancy raises a central question: does the breadth of association reflect a similar breadth of causal drivers, or does it mask a smaller number of upstream disturbances that shape the dominant disease trajectory?

In this manuscript, Crohn’s biology is organised into a depth-ordered five-node circuit linking microbial sensing, autophagy competence, epithelial stress resilience, inflammatory amplification, and barrier integrity. The proposed directional flow is NOD2 → ATG16L1/IRGM → XBP1 → IL23R → MUC2, with feedback from barrier compromise back to innate sensing. Rather than presenting this as a narrative synthesis, the framework is supported by explicit quantitative prioritisation and edge-confidence scoring.

Node leverage is derived from multiple real-data axes, including human genetic support, literature convergence, protein-interaction networks, and ileal single-cell disease activity profiles. Directed relationships between nodes are evaluated using a structured evidence matrix and accompanied by prespecified falsification tests. This approach allows the model to distinguish between high-confidence proximal links and inference-dominant middle or distal edges that require direct perturbational validation.

The analysis places NOD2 at the strongest upstream position in aggregate leverage, while IL23R remains the most robust translational anchor given established therapeutic efficacy of IL-23 pathway inhibition. XBP1 and MUC2 show comparatively strong epithelial disease-state activity despite more modest genetic anchoring, reinforcing their roles as tissue-state and feedback nodes rather than primary initiators.

Importantly, the framework is explicitly bounded to ileal-predominant disease and does not claim universal applicability across all Crohn’s phenotypes. Alternative orderings and weighting assumptions are stress-tested to evaluate robustness, and the model is designed to be revised if perturbational data fail to support its predictions.

This version supersedes the earlier conceptual draft by introducing quantitative scoring, formal evidence grading, ordering sensitivity analysis, and explicit failure criteria. The goal is not to declare a definitive causal chain, but to provide an auditable scaffold that clarifies depth of influence, prioritises experimental validation, and supports rational translational hypothesis testing.