From single-pathway cascade to network pathophysiology: how genetically engineered mouse models reshaped our...

Pancreatitis remains a common gastrointestinal disease with no mechanism-specific therapy. Its pathogenesis was long explained by the autodigestion hypothesis: a single linear cascade initiated by premature intra-acinar trypsinogen activation. This trypsin-centric view cannot account for the amplification of inflammation, the switching of cell-death modality, the acute-to-chronic transition, organ failure, or the chronic pancreatitis caused by protein-misfolding genes that act outside the protease system. Here we synthesize evidence from genetically engineered mouse models (GEMMs); their loss- and gain-of-function designs, spanning global and conditional knockouts, knock-ins, and CRISPR/Cas9 editing, can separate molecular events that are necessary, sufficient, or merely correlated. These models show, for example, that enhanced trypsinogen autoactivation alone is sufficient for spontaneous disease, that NF-κB-driven inflammation can be initiated independently of trypsin, and that loss of autophagy alone disrupts acinar homeostasis. On this basis we reframe pancreatitis as a network pathophysiology with two coupled layers. First, acinar homeostasis is maintained by several semi-independent control modules (digestive-enzyme safety, endoplasmic reticulum protein quality control, the autophagy-lysosome system, the calcium-mitochondria axis, and NF-κB signaling), so that disease can be initiated at any of several independent nodes. Second, the injured acinar cell dies through an interconnected, switchable network of regulated cell death (apoptosis, necroptosis, pyroptosis, ferroptosis, and PANoptosis), sharing nodes such as caspase-8, RIPK3, and GPX4, whose configuration, more than the initiating trigger, determines severity. This framework recasts the heterogeneity of human pancreatitis as different entry points into one network and argues for therapies directed at shared nodes rather than single pathways.