Rate of autoactivation determines pancreatitis phenotype in trypsinogen mutant mice

The digestive protease trypsin plays a central role in the onset and progression of pancreatitis [1]. Ectopic conversion of the inactive precursor trypsinogen to active trypsin in the pancreas is an early event in disease development. Biochemically, trypsinogen is activated by limited proteolysis of its activation peptide that can be cleaved by three different enzymes: enteropeptidase (enterokinase), trypsin (autoactivation), and cathepsin B (CTSB) (Figure 1A). Since enteropeptidase is not expressed in the pancreas, pathological intrapancreatic trypsinogen activation can occur only through autoactivation or by CTSB-mediated activation. Inborn mutations of human cationic trypsinogen (encoded by the serine protease 1, PRSS1, gene) accelerate trypsinogen autoactivation and cause hereditary chronic pancreatitis (CP) [2]. In contrast, human genetic studies and functional analysis of trypsinogen mutants failed to substantiate a pathogenic role for CTSB-mediated trypsinogen activation in CP [references in 3].

Figure 1.

Rate of autoactivation determines pancreatitis phenotype in trypsinogen mutant mice. A, Autoactivation of wild-type mouse cationic (T7) trypsinogen and mutants D23A, D22N,K24R, and K24R in vitro in 1 mM calcium and 0.1 M NaCl (pH 8.0). Mean values (n=3) with the standard deviation (SD) are shown. The amino-acid sequence of the T7 trypsinogen activation peptide and the location of mutations D22N and K24R (in red, underlined) are also indicated. B,C,D, Spontaneous, progressive pancreatitis in homozygous T7D22N,K24R mice (T7D22N,K24R^hom). Mice were euthanized at the indicated times and the pancreas was analyzed. B, Representative hematoxylin-eosin stained pancreas sections. The scale bar is 100 μm. C, Pancreas weight. The number of C57BL/6N and T7D22N,K24R^hom mice analyzed at the indicated time points were 11, 12, 10, 11, 15, 10, and 19, 20, 17, 13, 15, 7, respectively. D, Histological scoring for areas of acute and chronic pancreatitis, expressed as percent of the total tissue area. Each mouse is represented by two symbols; white indicates acute pancreatitis, black designates chronic pancreatitis. Areas of acute disease were identified as having relatively normal acini separated by edema with inflammatory cells evident in the tissue spaces. Areas of chronic pancreatitis were defined by acinar atrophy. The number of T7D22N,K24R^hom mice analyzed at the indicated time points were 10, 20, 17, 13, 15, and 7. C,D, Individual data points were graphed. Trend lines connecting the mean values at given time points illustrate the direction of change. E, Severe cerulein-induced acute pancreatitis and progression to chronic pancreatitis in heterozygous T7D22N,K24R mice (T7D22N,K24R^het). Trypsinogen mutant and C57BL/6N mice received 10 hourly injections of cerulein or saline. Mice were euthanized an hour (day 1) or a week (day 8) after the last injection, as indicated. Representative hematoxylin-eosin stained pancreas sections, and changes in pancreas weight are shown. Individual data points were graphed with the mean and SD indicated. The number of C57BL/6N and heterozygous T7D22N,K24R mice analyzed for pancreas weight were as follows. Cerulein: 13 and 12 (day 1), 6 and 10 (day 8), saline: 6 and 7 (day 1), 2 and 2 (day 8), respectively. See Supplementary Material for further experimental details. The scale bar is 100 μm.

The pathogenic role of increased trypsinogen autoactivation in pancreatitis has been recently confirmed by two mouse models carrying mutations in mouse cationic trypsinogen (isoform T7) [4–6]. The T7D23A and T7K24R mice harbor mutations D23A and K24R in the trypsinogen activation peptide that directly stimulate autoactivation by 50- and 5-fold, respectively. In contrast, these mutations do not increase CTSB-mediated trypsinogen activation. T7D23A and T7K24R mice exhibit strikingly different phenotypes with respect to pancreatic disease. Thus, heterozygous T7D23A mice develop spontaneous, early-onset acute pancreatitis (AP) with rapid progression to CP [4]. Conversely, homozygous T7K24R mice do not have spontaneous pancreatitis, but exhibit more severe disease than C57BL/6N control mice when pancreatitis is induced experimentally by repeated cerulein injections [5]. Furthermore, after the acute episode, T7K24R mice continue to progress to CP while control C57BL/6N mice recover quickly [6]. These properties of the T7D23A and T7K24R mice led us to postulate that the rate of autoactivation of the mutant trypsinogens is the key determinant of the pancreatitis phenotype.

To test this hypothesis, here we engineered a novel trypsinogen mutant, which exhibits an accelerated rate of autoactivation that lies between those observed previously with mutations D23A and K24R. Using recombinant mouse T7 trypsinogens, we found that the double-mutant D22N,K24R autoactivated at a 13-fold increased rate relative to wild-type T7 (Figure 1A), while CTSB-mediated activation was slightly decreased (not shown). We generated a new T7D22N,K24R mouse strain which carries these mutations in the activation peptide of T7 trypsinogen (Figure 1A). When analyzed for spontaneous pancreatitis and severity of cerulein-induced pancreatitis, we found that homozygous and heterozygous T7D22N,K24R mice exhibited phenotypes that resembled those of the T7D23A and T7K24R mice, respectively. Thus, homozygous T7D22N,K24R mice developed spontaneous, progressive pancreatitis, starting around 6 weeks of age (Figures 1B,C,D, Supplementary Figure S1). In a subset of mice, signs of AP (edema, inflammatory cells, elevated plasma amylase) were apparent at the age of 6–8 weeks. Starting at the same time, mice developed progressive atrophic CP with fibrosis and eventual adipose replacement. Elevated intrapancreatic trypsin activity, relative to C57BL/6N controls, was apparent throughout the disease course. This spontaneous pancreatitis phenotype was highly reminiscent of that of T7D23A mice [4], even though the age of onset and rate of progression was delayed.

In contrast to homozygous T7D22N,K24R mice, heterozygous mutants did not show signs of spontaneous pancreatic disease. However, when AP was induced by cerulein, higher intrapancreatic trypsin activity, and more severe pathology (as judged by plasma amylase activity, pancreas edema, and histology) was observed in heterozygous T7D22N,K24R mice relative to C57BL/6N controls (Figure 1E, Supplementary Figures S2A,B,C). Remarkably, after an episode of cerulein-induced AP, heterozygous T7D22N,K24R mice progressed to CP, in a manner that was previously observed with homozygous T7K24R mice (Figure 1E, Supplementary Figure S2D) [6]. Seven days after the acute episode (on day 8), the pancreas of heterozygous T7D22N,K24R mice was highly atrophic with acinar cell ablation, dilated ducts, fibrosis, and inflammatory cells, whereas C57BL/6N mice showed complete histological restitution. Taken together, the experimental pancreatitis of heterozygous T7D22N,K24R mice phenocopies that of homozygous T7K24R mice both in terms of severity of AP and progression to CP.

The observations that the novel T7D22N,K24R trypsinogen mutant mice replicate the two previously described pancreatitis phenotypes as a function of zygosity, convincingly demonstrate that increased trypsinogen autoactivation can cause various manifestations of pancreatitis and the rate of autoactivation is the key determinant when and how pathology develops. If we take into consideration the number of mutant trypsinogen alleles (1× or 2×), and the relative rate of autoactivation (50-,13-, and 5-fold increased), we can rank the various strains studied with respect to their “trypsinogen autoactivation potential”: heterozygous T7D23A (1×50) > homozygous T7D22N,K24R (2×13=26) > heterozygous T7D22N,K24R (1×13) ≅ homozygous T7K24R (2×5=10). Mouse strains with a high propensity for trypsinogen autoactivation develop spontaneous, progressive pancreatitis. In contrast, mice with lower autoactivation potential will not have spontaneous disease, yet they still exhibit heightened sensitivity to experimentally induced AP and a tendency for progression to CP after an acute episode.

The findings presented here have important implications with respect to human CP associated with genetic changes. Mutations that increase the rate of trypsinogen autoactivation above a certain threshold are expected to be highly penetrant and cause hereditary CP, while mutations with a lesser effect should be considered only risk factors for CP development. Human cationic trypsinogen has a stronger propensity for autoactivation than its murine counterpart, suggesting that even smaller increases than those described here may be sufficient to reach the critical limit required to trigger pancreatitis. Recently generated mouse models with human trypsinogen transgenes can help to decipher this quantitative relationship [7–9].

Finally, the novel T7D22N,K24R strain is an excellent tool for preclinical testing of drug candidates against trypsin-dependent pancreatitis. The fact that both spontaneous and experimentally induced pancreatitis can be studied in the same model is convenient, cost-effective, and accelerates the preclinical discovery and development process.

In summary, here we demonstrated that the pancreatitis phenotype of mice carrying trypsinogen mutations is determined by the rate of autoactivation of the trypsinogen mutants. High autoactivation rates are associated with spontaneous, progressive disease whereas lower autoactivation rates cause sensitization to experimental pancreatitis with increased severity and subsequent progression to CP.

Data and materials:

All materials are available for research purposes upon request. All data are included in the manuscript.