Genetically Engineered Mouse Models Shine New Light on Decades-old Story of Trypsin in Pancreatitis

The idea that pancreatitis—a common, painful, and life-threatening disease^1,2—is caused by the pancreas digesting itself is about 130 years old,³ and the foremost villain in this process is believed to be trypsin, a serine protease discovered even earlier. The problem is that exocrine pancreatic (acinar) cells synthesize not trypsin per se, but trypsinogen, its inactive precursor (zymogen). In normal physiology, trypsinogen is converted to trypsin (activated) only in the duodenum after being secreted in response to a meal. In addition, acinar cells have mechanisms protecting against active trypsin and premature zymogen activation, such as an endogenous trypsin inhibitor. Therefore, how trypsinogen is activated inside the pancreas and what exactly is the pathological role of trypsin, particularly in the disease onset, have for decades been central questions in the mechanism of pancreatitis.⁴

Two mechanisms have been established for trypsinogen activation: (1) autoactivation (ie, that trypsin can generate itself by cleaving trypsinogen) and (2) trypsinogen cleavage by the endo/lysosomal protease cathepsin B (CTSB). Both were discovered in vitro biochemically^5,6 and subsequently examined in experimental and genetic in vivo rodent models of pancreatitis and ex vivo, on isolated acinar cells. Despite hundreds of these studies, the relative significance of the 2 pathways, their detailed mechanisms (eg, where exactly CTSB meets trypsinogen inside the acinar cell), and their relevance to human disease remain hotly debated.⁴ Evidence that the disease is associated with trypsinogen activation is the presence in the blood and urine of pancreatitis patients of trypsinogen activation peptide that is cleaved off the trypsinogen molecule during its conversion to trypsin.⁷ Furthermore, trypsinogen activation peptide levels correlate with disease severity. Another link was the 1996 discovery⁸ that the autosomal-dominant condition of hereditary pancreatitis is associated with a point mutation (p.R122H) in human cationic trypsinogen PRSS1. However, hereditary pancreatitis represents a tiny portion of the disease, and not all genetic mutations in several proteins linked so far to pancreatitis result in increased trypsin activity.^4,9,10

There have been twists and turns in the quest to elucidate the roles of autoactivation versus CTSB-mediated trypsinogen activation. For example, a selective CTSB inhibitor prevented intrapancreatic trypsin increases in ex vivo and in vivo acute pancreatitis (AP) models, including the classical model induced with high-dose cerulein, a cholecystokinin analog (CER-AP).^11,12 This ameliorated some CER-AP responses, but parameters of inflammation were largely unaffected.¹² Similarly, a total body CTSB knockout greatly (but not completely) abolished cerulein-induced trypsinogen activation and reduced necrosis without affecting the inflammatory response.¹³ (Of note,CTSB also regulates multiple trypsin-unrelated pathways.¹⁴) These results strengthened the CTSB-centered paradigm of trypsinogen activation,¹¹ but at the same time questioned the centrality of this pathway in pancreatitis pathogenesis.

The first genetically engineered mouse model (GEMM) targeting trypsinogen was the knockout of its mouse cationic T7 isoform.¹⁵ The cerulein-induced increase in intrapancreatic trypsin was abrogated in these mice, indicating that T7 is the isoform involved; acinar cell necrosis was decreased, but there was no effect on histopathology and inflammation. The objective of more recently developed GEMMs^16–19 has been to enhance trypsinogen autoactivation, either through mutations in T7 mimicking those in hereditary pancreatitis or by expressing human trypsinogens, including knock-in of the most common PRSS1 p.R122H human mutations. The first type of GEMMs displayed higher rates of autoactivation.^16,18 Increases in basal pancreatic trypsin activity were, however, reported only for unnatural T7 mutant with a 50-fold higher autoactivation rate, and only this mutant developed spontaneous pancreatitis.¹⁶ Similarly, spontaneous pancreatitis did not develop in mice expressing PRSS1 p.R122H,¹⁷ but only with the combined expression of PRSS1 p.R122H and PRSS2 human trypsinogens.¹⁹ Importantly, however, the enhanced trypsinogen autoactivation worsened CER-AP in all these GEMMs^16–19 and/or accelerated its progression to chronic disease.

In this issue of Gastroenterology, the study by Geisz et al²⁰ introduced new GEMMs carrying trypsinogen mutations that block autoactivation but preserve or even enhance CTSB-mediated activation and examined their effects in the CER-AP model. Abolishing autoactivation did not decrease trypsin activity elicited by cerulein, whereas enhancing CTSB-mediated trypsinogen activation increased pancreatic trypsin in CER-AP by approximately 3-fold, compared with wild type. The authors also generated a novel total body CTSB knockout, in which cerulein-induced trypsin activity was completely abrogated. The results show that trypsinogen activation in CER-AP is exclusively mediated by CTSB, largely reinforcing—in a new, elegant way—previous reports.^12,13 However, the major, less expected, and more consequential finding is that changes in the amount of CTSB-mediated intrapancreatic trypsin in these GEMMs had no effect on CER-AP responses measured in the study, such as serum hyperamylasemia, pancreatic necrosis, and infiltration of neutrophils and macrophages.

The findings of Geisz et al²⁰ match those of another recent study²¹ that generated pancreas-specific knockout of CTSB and similarly found no effect on a set of CER-AP responses. The two studies establish that (1) autoactivation does not mediate intrapancreatic trypsin increase in CER-AP, and (2) blocking or increasing CTSB-mediated trypsin activity has no effect on CER-AP responses. The first conclusion is unequivocal; a caveat to the second is that the authors only measured a subset of disease parameters at a 1-time point. The results^20,21 imply that CTSB-mediated increase in intrapancreatic trypsin—at least in this particular model—is not pathogenic. In contrast, enhancing trypsinogen autoactivation by genetically modifying T7 or expressing human trypsinogens aggravates the severity of the CER-AP model in corresponding GEMMs.^16–19 A possible explanation (as the authors speculate) is the potentially different sites of intrapancreatic trypsinogen activation: within acinar cells when mediated by CTSB versus in the interstitial space in the case of autoactivation.

The findings from Geisz et al²⁰ evoke additional questions and suggest new directions for research into the role of trypsin in pancreatitis—and, more generally, into the disease pathogenic mechanism. A more detailed characterization of CER-AP responses is needed in these GEMMs, which might reveal distinct parameters affected by the varying levels of trypsin activity (as found with pharmacologic CTSB inhibition¹²). It would be informative to measure the level and distribution of trypsinogen activation peptide as a complementary indicator of trypsinogen activation and to perform ex vivo studies in acinar cells isolated from the various GEMMs. An intriguing question is what would happen in a genetic mouse model in which the CTSB-mediated mechanism of trypsinogen activation is selectively abolished, but autoactivation is preserved or even enhanced. And perhaps the most revealing will be the application of the GEMMs developed by Geisz et al²⁰ to other dissimilar AP models, such as induced with L-arginine, a choline-deficient ethionine-supplemented diet, or ethanol.

The mouse models in the present study (and future related GEMMs) will serve as a valuable resource to unravel the decades-old enigma of trypsin in pancreatitis and could open new therapeutic venues for this debilitating disease.