Evolutionary expansion of polyaspartate motif in the activation peptide of mouse cationic trypsinogen limits autoactivation and protects against pancreatitis

Abstract

The activation peptide of mammalian trypsinogens typically contains a tetra-aspartate motif (positions P2–P5 in Schechter–Berger numbering) that inhibits autoactivation and facilitates activation by enteropeptidase. This evolutionary mechanism protects the pancreas from premature trypsinogen activation while allowing physiological activation in the gut lumen. Inborn mutations that disrupt the tetra-aspartate motif cause hereditary pancreatitis in humans. A subset of trypsinogen paralogs, including the mouse cationic trypsinogen (isoform T7), harbor an extended penta-aspartate motif (P2–P6) in their activation peptide. Here, we demonstrate that deletion of the extra P6 aspartate residue (D23del) increased the autoactivation of T7 trypsinogen threefold. Mutagenesis of the P6 position in wild-type T7 trypsinogen revealed that bulky hydrophobic side chains are preferred for maximal autoactivation, and deletion-induced shift of the P7 Leu to P6 explains the autoactivation increase in the D23del mutant. Accordingly, removal of the P6 Leu by NH₂-terminal truncation with chymotrypsin C reduced the autoactivation of the D23del mutant. Homozygous T7D23del mice carrying the D23del mutation did not develop spontaneous pancreatitis and severity of cerulein-induced acute pancreatitis was comparable with that of C57BL/6N controls. However, sustained stimulation with cerulein resulted in markedly increased histological damage in T7D23del mice relative to C57BL/6N mice. Furthermore, when the T7D23del allele was crossed to a chymotrypsin-deficient background, the double-mutant mice developed spontaneous pancreatitis at an early age. Taken together, the observations argue that evolutionary expansion of the polyaspartate motif in mouse cationic trypsinogen contributes to the natural defenses against pancreatitis and validate the role of the P6 position in autoactivation control of mammalian trypsinogens.

NEW & NOTEWORTHY Unwanted autoactivation of the digestive protease trypsinogen can result in pancreatitis. The trypsinogen activation peptide contains a polyaspartate motif that suppresses autoactivation. This study demonstrates that evolutionary expansion of these aspartate residues in mouse cationic trypsinogen further inhibits autoactivation and enhances protection against pancreatitis.

INTRODUCTION

The pancreas releases digestive proteases to the duodenum as inactive precursors, which become activated through limited proteolysis. Enteropeptidase (enterokinase) activates trypsinogens to trypsin which, in turn, converts chymotrypsinogens, proelastases, and procarboxypeptidases to their active forms (1). Full activation of procarboxypeptidases A1 and A2 requires further digestion of their propeptide by chymotrypsin C (2). Trypsinogens can also undergo autoactivation, initiated by their intrinsic zymogen activity and further amplified by the resulting trypsin activity. Although activation of trypsinogen by trypsin (i.e., autoactivation) is remarkably inefficient when compared with enteropeptidase-mediated activation, it is of pathological significance as premature, intrapancreatic autoactivation of trypsinogen can drive onset and progression of pancreatitis (3). Inborn mutations that stimulate autoactivation of trypsinogen are associated with hereditary pancreatitis and mouse models harboring similar mutations confirmed the causal relationship (3–5).

Activation of trypsinogen to trypsin requires proteolytic removal of the so-called activation peptide by cleaving a conserved Lys-Ile peptide bond (positions P1–P1′ in the Schechter-Berger numbering of protease substrates) (6). The activation peptide in mammalian trypsinogens is typically eight amino acids long and contains a tetra-Asp motif before the P1 Lys activation site (positions P2–P5) (7). A large body of work investigated the significance of the tetra-Asp motif and found that it serves as a recognition sequence for enteropeptidase and it inhibits autoactivation of trypsinogen (8 and references therein). Millimolar concentrations of calcium relieve this inhibition and accelerate autoactivation by neutralizing the negative charge of the Asp residues (9–11). We previously characterized the role of the four Asp residues (D19–D22) within the activation peptide of human cationic trypsinogen and demonstrated that in this isoform the Asp residues per se are not required for enteropeptidase-mediated activation while they are essential for autoactivation control (8). Thus, individual mutations D19A, D20A, and D21A stimulated autoactivation two- to threefold, whereas the combination of these mutations caused a 13-fold increase. Mutation D22A caused a dramatic 66-fold increase in autoactivation and complete “Ala-shave” of the tetra-Asp motif had a more than 500-fold effect. We also found that inhibition of autoactivation is dependent on the Asp218 exosite, which interacts electrostatically with the tetra-Asp motif of the activation peptide. Asp218 is unique to human cationic trypsinogen indicating that trypsinogens may evolve isoform-specific mechanisms to control autoactivation.

Curiously, the tetra-Asp motif is extended by an Asp residue in mouse and rat cationic trypsinogens (12). A genomic database search reveals that similar penta-Asp motifs exist not only in rodents but also in the trypsinogens of at least 23 other species, including various bats, the rabbit, and some exotic animals such as the koala, platypus, giant panda, and the Tasmanian devil. Here, we set out to investigate the biological significance of this evolutionary variation with respect to trypsinogen autoactivation control and protection against pancreatitis. In addition to the obligatory biochemical studies, we also generated and characterized a new mouse model in which the penta-Asp motif in cationic trypsinogen was shortened to a tetra-Asp sequence.

EXPERIMENTAL PROCEDURES

Materials

Protease substrates Z-Gly-Pro-Arg-p-nitroanilide (GPR-pNA, Cat. No. 4000768), Z-Gly-Pro-Arg-7-amino-4-methylcoumarin (GPR-AMC, Cat. No. 4002047), and Suc-Ala-Ala-Pro-Phe-7-amino-4-methylcoumarin (AAPF-AMC, Cat. No. 4012873) were obtained from Bachem (Torrance, CA). Cerulein (Cat. No. C9026) was purchased from MilliporeSigma. Cathepsin B (CTSB, Cat. No. 219362-50UG) and cathepsin L (CTSL, Cat. No. 219402-25UG) purified from human liver were purchased from EMD Millipore (Temecula, CA).

Accession Numbers and Nomenclature

NC_000072.6, Mus musculus strain C57BL/6J chromosome 6, GRCm38.p6 C57BL/6J; NM_023333.4, Mus musculus RIKEN cDNA 2210010C04 gene (2210010C04Rik) mRNA, encoding mouse cationic trypsinogen (isoform T7). Amino acid residues in trypsinogen were numbered starting from the initiator methionine. Because of an additional Asp residue in the activation peptide, numbering in T7 trypsinogen is shifted by one relative to human trypsinogens (12).

Mutagenesis, Expression, and Purification of Trypsinogen

The pTrapT7 plasmid harboring the coding sequence for mouse T7 trypsinogen was constructed previously (12). Mutations were introduced by overlap extension PCR mutagenesis. Wild-type and mutant proteins were expressed in Escherichia coli BL21 (DE3), the inclusion bodies were isolated and subjected to in vitro refolding (13). Correctly folded trypsinogen was purified by ecotin affinity-chromatography (13, 14). Concentration of T7 trypsinogen solutions was calculated from the UV absorbance at 280 nm using the extinction coefficient 39,140 M⁻¹·cm⁻¹. The experiments studying the effects of cathepsins used trypsinogen preparations expressed in E. coli LG-3 from the pTrapT7-intein-mouse-T7 plasmid that contains the coding sequence for T7 trypsinogen fused to an NH₂-terminal self-splicing mini-intein (4, 13, 15).

Trypsinogen Autoactivation

Trypsinogen (2 μM) was incubated with 10 nM initial trypsin in 0.1 M Tris·HCl (pH 8.0), 1 mM CaCl₂, and 0.05% Tween 20 (final concentrations) in 100 μL final volume, at 37°C. At given time points, aliquots (1.5 µL) were removed, mixed with 48.5 µL assay buffer [0.1 M Tris·HCl (pH 8.0), 1 mM CaCl₂, and 0.05% Tween 20], and trypsin activity was measured after addition of 150 µL GPR-pNA substrate (200 μM solution in assay buffer). Trypsin activity was expressed as percent of the potential full activity measured after activation with enteropeptidase.

NH₂-Terminal Truncation of Trypsinogen by CTRC

Chymotrypsin C (CTRC)-mediated cleavage of the activation peptide in the D23del T7 trypsinogen mutant was analyzed by incubating trypsinogen (2 µM) with 25 nM mouse CTRC in 0.1 M Tris·HCl (pH 8.0). To prevent trypsinogen autoactivation, the reaction also contained 25 nM human serine protease inhibitor Kazal type 1 (SPINK1) trypsin inhibitor. At the indicated times, trypsinogen (100 µL) was precipitated with 10% trichloroacetic acid (final concentration) and intact and processed forms were resolved by SDS-PAGE under nonreducing conditions followed by Coomassie Blue staining.

Trypsinogen Activation and Inactivation by Cathepsins

Experiments to characterize activation of trypsinogen by cathepsin B (CTSB) and inactivation by cathepsin L (CTSL) were carried out exactly as reported in our recent publication (15). Before use, cathepsin preparations were incubated with 0.5 mM dithiothreitol (DTT) for 30 min, to fully reduce the active site cysteine.

SDS-PAGE Analysis

Trypsinogen (100 μL) was precipitated with 10% trichloroacetic acid (final concentration) and centrifuged for 10 min at 20,000 g, and the pellet was dissolved in 20 μL Laemmli sample buffer containing 100 mM DTT, unless indicated otherwise. Samples were heat-denatured at 95°C for 5 min and electrophoresed on 15% SDS-polyacrylamide minigels. Bands were visualized by Coomassie Blue R-250 staining.

Animal Studies Protocol Approval

Animal experiments were performed at Boston University and the University of California Los Angeles (UCLA) with the approval and oversight of the Institutional Animal Care and Use Committee (IACUC) of Boston University and the Animal Research Committee (ARC) at UCLA, including protocol review and postapproval monitoring. The animal care programs at these institutions are managed in full compliance with the United States Animal Welfare Act, the United States Department of Agriculture Animal Welfare Regulations, the United States Public Health Service Policy on Humane Care and Use of Laboratory Animals, and the National Research Council’s Guide for the Care and Use of Laboratory Animals. Boston University and UCLA have approved Animal Welfare Assurance statements (A3316-01 and A3196-01, respectively) on file with the United States Public Health Service, National Institutes of Health, Office of Laboratory Animal Welfare. Both institutions are accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC).

Mouse Strains

The novel T7D23del line was generated by introducing the c.67_69del (D23del) mutation into the T7 trypsinogen gene via homologous recombination in C57BL/6 embryonic stem (ES) cells (Cyagen, Santa Clara, CA). The ∼3.8-kb gene encoding T7 trypsinogen comprises five exons, and it is located on chromosome 6 in mice. The targeting vector contained the T7 trypsinogen gene with the D23del mutation in exon 2 and the neomycin resistance gene flanked by loxP sites in intron 1 (appendix Fig. A1). Recombinant ES cell clones were identified by long-range PCR and confirmed by Southern blotting. ES cells were injected into mouse embryos (blastocysts), which were implanted into pseudopregnant females. The resulting chimeras were bred with C57BL/6N mice to achieve germline transmission of the mutant allele. To remove the neomycin cassette from this F1 generation, the mice were bred with a Cre-deleter strain that expresses the Cre recombinase in the mouse embryo. The final T7D23del allele contained the D23del mutation in exon 2 and a 127 nt residual sequence in intron 1 including a single loxP site (appendix Fig. A2). T7D23del mice were maintained in the homozygous state. Generation and properties of the Ctrb1-del knockout mice deficient in chymotrypsin B1 (CTRB1) were described previously (16). In the present study, Ctrb1-del mice were crossed with T7D23del mice to produce a double homozygous T7D23del × Ctrb1-del strain. C57BL/6N mice were obtained from Charles River Laboratories (Wilmington, MA) or were bred from the same stock. The number of mice used in each experiment is shown in the appendices. Both male and female animals were studied. Mice used for cerulein injection experiments were 11- to 12-wk old and weighed typically 24–25 g (males) and 19–20 g (females).

Genotyping

To genotype mice for the T7D23del allele, exon 2 with flanking intronic sequences was amplified from genomic DNA using primers given in appendix Fig. A2. C57BL/6N and T7D23del mice yielded 630-bp and 754-bp long products, respectively. The Ctrb1-del allele was genotyped as described previously (16).

Quantitative Reverse Transcription PCR

RNA was isolated from 30 mg of pancreas tissue and reverse transcribed (17). Levels of T7 trypsinogen mRNA were measured by quantitative real-time PCR, as described recently (5). Relative expression levels were calculated with the comparative cycle threshold method (ΔΔCT method).

Western Blot Analysis

Pancreas tissue (40 mg) was homogenized in 400 µL phosphate-buffered saline (pH 7.4) supplemented with 4 µL Halt Protease and Phosphatase Inhibitor Cocktail (100×, Cat. No. 78440, Thermo Fisher Scientific). The homogenate was clarified by centrifugation (10 min, 13,500 rpm, 4°C), and 30 µg of total protein was resolved on 15% SDS-PAGE and transferred to a PVDF membrane. To detect T7 trypsinogen, a custom-made rabbit polyclonal antibody was used at 1:10,000 dilution (4). Mouse ERK1/2 was detected with a rabbit monoclonal antibody (Cat. No. 4695, Cell Signaling Technology) used at 1:500 dilution. HRP-conjugated goat anti-rabbit polyclonal IgG (Cat. No. 31460, Thermo Fisher Scientific) was used as secondary antibody at 1:10,000 dilution after the T7 trypsinogen primary antibody and at 1:20,000 dilution after the ERK1/2 primary antibody.

Cerulein-Induced Pancreatitis in Mice

Acute pancreatitis in C57BL/6N and T7D23del mice was induced by 10 hourly intraperitoneal injections of cerulein (50 µg/kg body wt), as described previously (16, 17). Control mice were given normal saline injections. Mice were euthanized 1 h after the last injection, and the pancreas and blood were harvested. For sustained stimulation with cerulein, hourly injections were performed eight times per day on two consecutive days. Controls were given normal saline injections. Mice were euthanized 3 days after the last injection. Histological analysis of the pancreas and measurement of plasma amylase activity, pancreatic water content, and myeloperoxidase (MPO) levels were carried out as described previously (5, 17).

Intrapancreatic Trypsin and Chymotrypsin Activation

Cerulein-induced intrapancreatic activation of trypsinogen and chymotrypsinogen was measured 30 min after a single injection of cerulein (50 µg/kg body wt), according to our recently published method (18). Control mice were given normal saline injections.

Statistical Analysis

Results of animal experiments were graphed as individual data points with the mean and standard deviation (SD) indicated. The difference of means between two groups was analyzed by unpaired t test. P < 0.05 was considered statistically significant.

RESULTS

Autoactivation of the D23del Mouse Cationic Trypsinogen Mutant

To assess the effect of the extended penta-Asp motif in the activation peptide on autoactivation, we deleted an Asp residue in the mouse cationic trypsinogen (isoform T7). The NH₂-terminal amino acid of mature, secreted T7 trypsinogen is Leu16. Because the sequence alignment starts at the NH₂-terminal end, the correct designation of the new mutant with the missing Asp is D23del (Fig. 1A). With respect to the P1–P1′ Lys24-Ile25 activation-site residues, the deleted Asp corresponds to position P6 (Fig. 1B). When measured at pH 8.0, in 1 mM calcium, autoactivation of the purified D23del mutant was accelerated at least threefold relative to wild-type T7 trypsinogen (Fig. 1C).

Figure 1.

Effect of activation peptide mutations on the autoactivation of T7 trypsinogen. A: the activation peptide sequences of wild-type T7 trypsinogen and mutant D23del. The position of the deleted Asp23 is indicated as a gap. The numbers indicate the amino acid positions counted from the initiator methionine of pretrypsinogen. B: alignment of the activation peptides of wild-type T7 trypsinogen and mutant D23del with respect to the activation site Lys24. The amino acid positions are denoted according to the Schechter-Berger nomenclature. C: autoactivation of wild-type T7 trypsinogen and mutant D23del at pH 8.0, in 1 mM CaCl₂. D: effect of mutations of Asp19 on the autoactivation of T7 trypsinogen. Experiments were performed as described in experimental procedures. Trypsin activity was expressed as percentage of the potential full activity.

The P6 Residue in the Trypsinogen Activation Peptide Controls Autoactivation

In the D23del mutant, the original P7 Leu residue is shifted to the P6 position (see Fig. 1B). To test the significance of the newly positioned P6 Leu in autoactivation, we mutated the P6 Asp residue in wild-type T7 trypsinogen to Leu, Phe, Ile, Val, Ala, and Thr. Remarkably, autoactivation of mutants D19L and D19F were increased to the same extent as seen with mutant D23del (Fig. 1D). Smaller but still significant increases were seen with mutants D19I and D19V, whereas mutants D19A and D19T autoactivated similarly to wild-type T7 trypsinogen. The results indicate that a hydrophobic residue is preferred at position P6 for maximal autoactivation and the presence of the P6 Leu fully explains the gain-of-function phenotype of the D23del mutant. The importance of the P6 Leu was further confirmed by proteolytic removal of the NH₂-terminal three amino acids from the D23del mutant by chymotrypsin C (CTRC). As shown previously for wild-type T7 trypsinogen (12), mouse CTRC rapidly cleaved the activation peptide after Leu18 resulting in an NH₂-terminally truncated trypsinogen (Fig. 2A). The small change in size can be visualized on SDS-PAGE under nonreducing conditions (Fig. 2B). As expected, removal of the P6 Leu by CTRC decreased the autoactivation of mutant D23del (Fig. 2C). As reported previously (see Figure 4D in Ref. 12), CTRC cleavage of the activation peptide in wild-type T7 trypsinogen had no impact on autoactivation, indicating that residues beyond the P6 position (i.e., P7, P8, and P9, see Fig. 1B) play no role in autoactivation control. Taken together, the results indicate that evolutionary expansion of the tetra-Asp motif in trypsinogens curbs autoactivation by introducing an Asp residue to position P6.

Figure 2.

NH₂-terminal truncation of the D23del T7 trypsinogen mutant by mouse chymotrypsin C (CTRC). A: the activation peptide sequence of the D23del T7 trypsinogen mutant before and after CTRC-mediated cleavage. B: analysis of CTRC-mediated NH₂-terminal truncation of D23del T7 trypsinogen mutant by nonreducing SDS-PAGE and Coomassie blue staining. C: effect of CTRC-mediated NH₂-terminal cleavage on the autoactivation of the D23del T7 trypsinogen mutant. Experiments were performed as described in experimental procedures. Similar experiments on wild-type T7 trypsinogen were published previously (see text for details).

Cathepsin-Mediated Processing of T7 Trypsinogen is Unaffected by the D23del Mutation

Trypsinogen may become activated intracellularly by cathepsin B (CTSB), which cleaves the same activation site (Lys24-Ile25 in T7 trypsinogen) as enteropeptidase and trypsin (15 and references therein). In contrast, cathepsin L (CTSL) inactivates trypsinogen through proteolytic cleavage of a peptide bond (Gly27-Gly28 in T7 trypsinogen) adjacent to the activation site (15 and references therein). We tested the effect of cathepsins and found that wild-type T7 trypsinogen and the D23del mutant were activated by CTSB (Fig. 3A) and degraded by CTSL (Fig. 3B) at similar rates. Therefore, we conclude, the effect of the D23del mutation is specific to autoactivation of trypsinogen and does not alter cathepsin-mediated cleavages.

Figure 3.

Activation by cathepsin B and inactivation by cathepsin L of wild-type and D23del mutant T7 trypsinogens. A: activation of trypsinogens by cathepsin B at pH 4.0. B: cleavage of trypsinogens by cathepsin L at pH 4.0, analyzed by reducing SDS-PAGE and Coomassie Blue staining. Experiments were carried out as described in experimental procedures.

Generation of the T7D23del Mouse Model

We were curious whether the identified evolutionary protective mechanism has a measurable biological impact in vivo. To answer this question, we introduced the D23del mutation into exon 2 of T7 trypsinogen in C57BL/6N mice, as described in experimental procedures (Fig. 4A). No phenotypic changes were apparent in homozygous T7D23del mice, which grew and bred normally. Expression of T7 trypsinogen was comparable in C57BL/6N and T7D23del mice, as judged by quantitative reverse-transcription PCR (Fig. 4B) and by Western blotting (Fig. 4C).

Figure 4.

Generation of the T7D23del mouse strain and expression of T7 trypsinogen. A: schematic representation of the wild-type T7 trypsinogen gene and the recombined T7D23del allele before and after excision of the neomycin cassette. Exons are shown as dark yellow boxes. The loxP sites are in red. Thick blue lines indicate the homology arms. See text for further details. B: expression of T7 trypsinogen mRNA in the pancreas of C57BL/6N and T7D23del mice. Quantitative reverse-transcription PCR was performed as described in experimental procedures. Individual data points were plotted with the means and SD indicated. The difference of means between two groups was analyzed by two-tailed unpaired t test. C: protein levels of T7 trypsinogen in the pancreas of C57BL/6N and T7D23del mice. Western blotting of pancreas homogenates was performed as described in experimental procedures. ERK1/2 was measured as loading control.

Cerulein-Induced Acute Pancreatitis in T7D23del Mice

We speculated that increased autoactivation of the D23del trypsinogen mutant might result in pancreatitis in T7D23del mice or increase their susceptibility to the disease. We observed no spontaneous pancreatitis in T7D23del mice followed up to 6 mo of age, as assessed by histology (not shown). We induced experimental acute pancreatitis in C57BL/6N and T7D23del mice by 10-hourly intraperitoneal injections of the secretagogue cerulein and euthanized the mice 1 h after the last injection. On hematoxylin-eosin-stained pancreas sections from the cerulein-treated mice, we observed the characteristic edema, inflammatory cell infiltration, and scattered acinar cell necrosis, which were absent in mice given saline (Fig. 5A). Histological scoring of multiple sections for edema (Fig. 5B), inflammatory cells (Fig. 5C), and acinar cell necrosis (Fig. 5D) revealed that pancreatitis responses in the two strains were comparable, with a trend for higher edema and more inflammatory infiltrates in T7D23del mice. Edema was also evaluated by measuring the pancreas weight and the pancreatic water content (Fig. 6). Both parameters increased significantly in response to cerulein treatment in the two strains. Again, a nonsignificant trend for higher pancreatic weight was evident in cerulein-treated T7D23del mice relative to C57BL/6N mice (Fig. 6A), but the difference disappeared when pancreas weight was normalized to body weight (Fig. 6B). No appreciable differences were seen in pancreatic water content in cerulein-treated T7D23del and C57BL/6N mice (Fig. 6C). As expected, marked elevations in plasma amylase activity and pancreatic myeloperoxidase (MPO) content were observed in mice given cerulein compared with saline-treated mice (Fig. 7). Plasma amylase activity in cerulein-treated T7D23del mice was significantly higher than in similarly treated C57BL/6N mice (Fig. 7A) while no difference was seen in pancreatic MPO levels (Fig. 7B). Taken together, the results indicate that the D23del mutation in T7 trypsinogen does not change the severity of cerulein-induced pancreatitis significantly. The higher plasma amylase levels in cerulein-treated T7D23del mice, however, suggest that this strain is slightly more prone to acinar cell damage than the C57BL/6N parent strain.

Figure 5.

Histology of cerulein-induced acute pancreatitis. C57BL/6N and T7D23del mice were given 10 hourly saline or cerulein injections, as indicated. Mice were euthanized 1 h after the last injection. A: hematoxylin-eosin-stained pancreas sections. The scale bars correspond to 100 µm (top four panels) and 50 µm (bottom two enlargements). Histology scoring of pancreas sections was performed for edema (B), inflammatory cell infiltration (C), and acinar cell necrosis (D). Individual data points were graphed with the means and SD indicated. The difference of means between two groups was analyzed by two-tailed unpaired t test. See experimental procedures for details.

Figure 6.

Pancreas weight and pancreatic water content in cerulein-induced pancreatitis. C57BL/6N and T7D23del mice were given 10 hourly saline or cerulein injections, as indicated. Mice were euthanized 1 h after the last injection. A: pancreas weight in milligram units. B: pancreas weight as percentage of body weight. C: pancreas water content expressed as percent of wet pancreas weight. Individual data points were graphed with the means and SD indicated. The difference of means between two groups was analyzed by two-tailed unpaired t test. See experimental procedures for details.

Figure 7.

Plasma amylase and pancreas myeloperoxidase (MPO) content in cerulein-induced pancreatitis. C57BL/6N and T7D23del mice were given 10 hourly saline or cerulein injections, as indicated. Mice were euthanized 1 h after the last injection. A: plasma amylase activity expressed in mOD/min units. B: pancreas MPO content. Individual data points were graphed with the means and SD indicated. The difference of means between two groups was analyzed by two-tailed unpaired t test. See experimental procedures for technical details.

Cerulein-Induced Intrapancreatic Protease Activation in T7D23del Mice

To examine whether the increased autoactivation of the D23del mutant translates to higher intrapancreatic protease activation during cerulein-induced pancreatitis, we measured trypsin and chymotrypsin activities from pancreas homogenates 30 min after a single cerulein or saline injection. Previous studies indicated that this is the optimal time point to assess intrapancreatic protease activation as the pancreas is still relatively intact and the acinar tissue is unaffected by necrosis or inflammatory cell infiltration. Both trypsin and chymotrypsin activities were markedly higher in the pancreas of mice given cerulein relative to saline-treated mice (Fig. 8). When the cerulein-treated groups were compared, a small but significant, increase in trypsin activity was observed in T7D23del mice versus C57BL/6N mice (Fig. 8A), whereas chymotrypsin activities were similar (Fig. 8B). The results indicate that cerulein-induced intrapancreatic protease activation is comparable in the two strains with a slightly stronger propensity for trypsin activation in T7D23del mice.

Figure 8.

Cerulein-induced intrapancreatic trypsin and chymotrypsin activity. C57BL/6N and T7D23del mice were given a single saline or cerulein injection, as indicated, and the mice were euthanized 30 min later. A: trypsin activity in pancreas homogenates. B: chymotrypsin activity in pancreas homogenates. Individual data points were graphed with the means and SD indicated. The difference of means between two groups was analyzed by two-tailed unpaired t test. See experimental procedures for details.

Cerulein-Induced Chronic Pancreatitis in T7D23del Mice

Given the small differences observed in intrapancreatic trypsin activation and acute pancreatitis severity in T7D23del mice, we reasoned that more sustained stimulation with cerulein may amplify the incipient pathology and result in overt disease. Therefore, we treated T7D23del and C57BL/6N mice with 8 hourly injections of cerulein on two consecutive days and euthanized the mice 3 days after the last injection. Relative to saline-treated mice, which showed no histological damage whatsoever, cerulein-treated C57BL/6N mice exhibited scattered, small lesions (Fig. 9A). In contrast, some cerulein-treated T7D23del mice developed severe disease, with almost complete acinar cell ablation, disorganized histological architecture, dilated ducts, fibrosis, and so-called pseudotubular complexes. These changes are characteristic of chronic pancreatitis. Histological scoring of multiple pancreas sections confirmed that C57BL/6N mice recovered almost completely from cerulein-induced pancreatitis within 3 days, while more than half of cerulein-treated T7D23del mice still showed highly significant pancreas injury at that time (Fig. 9B). Measurement of pancreas weight confirmed the histological findings (Fig. 10). Significant atrophy was evident in the cerulein-treated groups relative to saline-treated mice (Fig. 10A), but this effect disappeared in the C57BL/6N group when pancreas weight was normalized to body weight (Fig. 10B). Importantly, pancreas atrophy of cerulein-treated T7D23del mice was much more pronounced (Fig. 10A) and remained highly significant even after normalization to body weight (Fig. 10B). These observations demonstrate the heightened sensitivity of T7D23del mice to pancreatitis in the setting of sustained pathological stimuli.

Figure 9.

Histology of cerulein-induced chronic pancreatitis. C57BL/6N and T7D23del mice were given 8 hourly saline or cerulein injections on 2 consecutive days, as indicated. Mice were euthanized 3 days after the last injection. A: hematoxylin-eosin-stained pancreas sections. Severe disease, as shown, was observed in about half of T7D23del mice treated with cerulein. The scale bars correspond to 100 µm (top four panels) and 50 µm (bottom two enlargements). B: histology scoring for intact acini. Individual data points were graphed with the means and SD indicated. The difference of means between two groups was analyzed by two-tailed unpaired t test. See experimental procedures for details.

Figure 10.

Pancreas weight in cerulein-induced chronic pancreatitis. C57BL/6N and T7D23del mice were given 8 hourly saline or cerulein injections for 2 days, as indicated. Mice were euthanized 3 days after the last injection. A: pancreas weight in milligram units. B: pancreas weight as percentage of body weight. Individual data points were graphed with the means and SD indicated. The difference of means between two groups was analyzed by two-tailed unpaired t test. See experimental procedures for details.

Spontaneous Pancreatitis in T7D23del × Ctrb1-del Mice

The experiments indicate that the D23del mutation in mouse cationic trypsinogen is insufficient to elicit pancreatitis but it sensitizes the pancreas to noxious stimuli. To obtain additional evidence for this concept, we crossed the T7D23del mice with homozygous Ctrb1-del mice which are deficient in chymotrypsin B1 (CTRB1), the predominant mouse chymotrypsin isoform. CTRB1 protects the pancreas against pancreatitis by degrading trypsinogen. Deletion of the Ctrb1 gene sensitizes the pancreas for trypsin activation and increases the severity of cerulein-induced pancreatitis (16, 17). However, similarly to T7D23del, the Ctrb1-del genotype is insufficient to trigger pancreatitis in the absence of an injurious insult. We hypothesized that the combination of the T7D23del and Ctrb1-del alleles would exceed the anti-trypsin defenses of the pancreas and result in spontaneous pancreatitis. We analyzed the pancreas of 26 T7D23del × Ctrb1-del mice euthanized at 2 wk of age (Fig. 11). We observed various degrees of pancreas pathology; in some cases most of the pancreas looked nearly normal (Fig. 11A), whereas in other cases acute pancreatitis was evident with edema and massive inflammatory cell infiltration (Fig. 11B). Acute pancreatitis with chronic pancreatitis-like changes was also observed (Fig. 11C). Finally, in some mice, the pancreas showed chronic pancreatitis-like pathology with acinar cell atrophy, dilated ducts, fibrosis, and pseudotubular complexes (Fig. 11D). The results convincingly demonstrate that in the T7D23del × Ctrb1-del mice, increased autoactivation of the D23del trypsinogen mutant synergizes with the loss of protective chymotrypsin and drives pancreatitis onset and progression.

Figure 11.

Histology of the pancreas from 2-wk-old T7D23del × Ctrb1-del mice (n = 26). Hematoxylin-eosin-stained pancreas sections showing various degree of pathology. The scale bars correspond to 100 µm. A: relatively intact pancreas histology was seen on 15/26 sections. B: severe acute pancreatitis was seen on 5/26 sections. C: acute pancreatitis with chronic changes was seen on 3/26 sections. D: chronic pancreatitis was seen on 3/26 sections. See text for further details.

DISCUSSION

In the present study, we investigated the biological significance of the Asp expansion in the activation peptide of mouse cationic trypsinogen (isoform T7) with respect to autoactivation control and protection against pancreatitis. Autoactivation of trypsinogen can lead to unwanted trypsin activity inside the pancreas which causes pancreatitis (3). The trypsinogen activation peptide evolved to limit autoactivation through multiple acidic residues that precede the activation site Lys (7). Human trypsinogens and most other mammalian paralogs contain a tetra-Asp motif in their activation peptides, whereas this sequence is extended to a penta-Asp motif in mouse cationic trypsinogen (see Fig. 1A). The extra Asp residue occupies the P6 position in the activation peptide (see Fig. 1B). Using biochemical approaches, we found that the P6 Asp in the activation peptide limits trypsinogen autoactivation to a significant extent. Thus, deletion of the P6 Asp residue resulted in at least threefold increased autoactivation. In the D23del mutant, a Leu residue occupies the P6 position and mutagenesis of wild-type T7 trypsinogen revealed that a P6 Leu or other bulky, hydrophobic side chains (Phe, Ile, Val) stimulate autoactivation. The observations argue that evolutionary expansion of the tetra-Asp motif to a penta-Asp sequence in the mouse cationic trypsinogen activation peptide replaced the P6 Leu with the newly introduced Asp and thereby decreased autoactivation.

Interestingly, a number of trypsinogen paralogs contain bulky hydrophobic residues at P6 of the activation peptide, e.g., Phe in human trypsinogens, Ile in the ferret and guinea pig cationic trypsinogens, or Val in the dog and cat cationic trypsinogens (7, 19–22). Since hydrophobic residues at P6 increase trypsinogen autoactivation, the evolutionary selection of these amino acids seems counterintuitive. However, some of these trypsinogens have been found to exhibit extremely slow autoactivation (21, 22) suggesting that the relatively moderate stimulatory effect of the hydrophobic P6 position in the activation peptide may be inconsequential. In other cases, the human trypsinogens in particular, the presence of a P6 Phe residue was likely selected to allow regulation by CTRC. This phenomenon was best characterized in human cationic trypsinogen where CTRC-mediated cleavage of the activation peptide after the P6 Phe18 accelerates autoactivation about fourfold (23–26). Although proteolytic removal of the P6 Phe residue should inhibit autoactivation, the observed stimulation is due to the release of the inhibitory effect exerted by the unique Asp218 exosite (23). CTRC-mediated processing of human anionic trypsinogen at Phe18 causes the expected decrease in autoactivation, as this isoform contains a Tyr residue in place of Asp218 (23, 25). Paradoxically, the physiological significance of CTRC-mediated autoactivation stimulation is likely related to the main regulatory function of CTRC, which is trypsinogen degradation to protect the pancreas against harmful trypsin activity. During the initial phase of premature trypsinogen autoactivation, CTRC-mediated stimulation helps to generate the initial trypsin activity that can synergize with CTRC to promote trypsinogen degradation and thereby extinguish the development of further trypsin activity. Notably, inborn mutations such as A16V, P17T, and N29I stimulate CTRC-mediated processing of the activation peptide and thereby increase trypsinogen autoactivation to pathological levels and cause hereditary pancreatitis (23, 24, 26). Proteolytic regulation of the activation peptide in other mammalian paralogs has not been studied, but the example of human trypsinogens described earlier indicates that evolutionary driving forces for amino acid selection at the P6 position can be complex and difficult to predict.

In addition to the biochemical studies, we also examined whether autoactivation control by the Asp-expansion in the activation peptide of T7 trypsinogen had any biological significance. To investigate this problem, we introduced the D23del mutation into mice and characterized the resultant T7D23del strain with respect to sensitivity to pancreatitis relative to the C57BL/6N parent strain. The results demonstrated that the D23del mutation clearly sensitized mice to pancreas injury, which was best seen after sustained pathological stimulation or when the mice were crossed with the Ctrb1-del pathogenic allele. Under these conditions, the anti-trypsin defenses of the pancreas were overwhelmed resulting in pancreatitis. A caveat to this conclusion is that we did not measure trypsin activity directly from the pancreas of T7D23del × Ctrb1-del mice. Nonetheless, the observations argue that the extra Asp residue in the penta-Asp motif of T7 trypsinogen alters autoactivation in a biologically significant manner, which explains its evolutionary selection.

The concept that increased autoactivation of mouse cationic trypsinogen causes pancreatitis has been also validated by recently published mouse models. Thus, T7D23A mice, which carry the D23A mutation in the T7 activation peptide, develop spontaneous acute pancreatitis that rapidly progresses to chronic pancreatitis (4). The D23A mutation stimulates autoactivation 50-fold, and this drastic increase explains the aggressive pathology. In contrast, T7K24R mice, containing mutation K24R in T7 trypsinogen, do not develop spontaneous disease but exhibit markedly increased severity of cerulein-induced acute pancreatitis (5). Mutation K24R increases autoactivation only fivefold, which is consistent with the less pronounced disease phenotype compared with T7D23A mice. Thus, there appears to be a correlation between the extent of trypsinogen autoactivation and severity of pancreas pathology. The present study provides further support to this concept, as the D23del mutation increased autoactivation threefold, which is slightly lower than the effect of the K24R mutation. Accordingly, unlike T7K24R mice, the T7D23del strain did not develop severe acute pancreatitis when given cerulein and overt pancreatitis was only observed after prolonged injury.

In summary, the biochemical investigations and the mouse models discussed earlier conclusively identify intrapancreatic autoactivation of trypsinogen as a key pathogenic mechanism in the development of pancreatitis and argue that evolutionary mechanisms to control autoactivation are essential for pancreas health. The results presented here also confirm and extend the notion that targeting trypsin pharmacologically would benefit patients with hereditary pancreatitis caused by trypsinogen mutations.

GRANTS

This work was supported by the Department of Defense Grants W81XWH-14-1-0331 (PR130667) (to M.S.-T.), W81XWH-17-1-0333 (PR192583) (to Z.J.), and W81XWH-19-1-0003 (PR181046) (to A.G.), the National Institutes of Health (NIH) Grants R01 DK117809, R01 DK058088, and R01 DK082412 (to M.S.-T.), the American Gastroenterological Association (AGA) Research Scholar Award AGA2020-13-05 (to A.G.), an American Pancreatic Association Young Investigator in Pancreatitis grant award (to A.G.), and a Scholarship from the Rosztoczy Foundation (to B.C.N).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.S.-T. conceived and designed research; A.O., B.C.N., Z.J., A.G., D.M., and A.D. performed experiments; A.O., B.C.N., Z.J., A.G., D.M., A.D., and M.S.-T. analyzed data; A.O., B.C.N., Z.J., A.G., D.M., A.D., and M.S.-T. interpreted results of experiments; A.O., Z.J., A.D., and M.S.-T. prepared figures; M.S.-T. drafted manuscript; A.O., B.C.N., Z.J., A.G., D.M., A.D., and M.S.-T. edited and revised manuscript; A.O., B.C.N., Z.J., A.G., D.M., A.D., and M.S.-T. approved final version of manuscript.

APPENDIX

Figure A1 details the recombined T7D23del targeting sequence. Figure A2 presents the T7D23del allele after Cre-mediated deletion of the neomycin cassette.

Figure A1.

The recombined T7D23del targeting sequence. The 5′ homology arm starts at c.-4286 in the 5′ upstream region and the 3′ homology arm ends at c.595-136 in intron 4. The trypsinogen exons are highlighted in yellow, the translational start codon is in green and underlined. Deletion c.67_69del (D23del) in exon 2 is shown in red with strikethrough; the neomycin cassette in intron 1 (between nucleotides c.41-346 and c.41-345) is in blue and the loxP sites are in magenta. The translational start and stop codons of the neomycin gene are emboldened and underlined. Splice sites are underlined and italicized. Relative to the reference sequence, the T7D23del allele contains the c.-1110C>A and c.594+72_73insAvariations, indicated in gray.

Figure A2.

The T7D23del allele after Cre-mediated deletion of the neomycin cassette. The 5′ homology arm starts at c.-4286 in the 5′ upstream region and the 3′ homology arm ends at c.595-136 in intron 4. The trypsinogen exons are highlighted in yellow, the translational start codon is in green and underlined. Deletion c.67_69del (D23del) in exon 2 is shown in red with strikethrough; the 127 nt residual scar sequence in intron 1 (between nucleotides c.41-346 and c.41-345) is in blue and the remaining loxP site is in magenta. Splice sites are underlined and italicized. Relative to the reference sequence, the T7D23del allele contains the c.-1110C>A and c.594+72_73insAvariations, indicated in gray. The primers used for genotyping are also indicated: Forward primer: CTT GAA ACT AAC AGT GGA CCC T; Reverse primer: AAC TGT GCA CAT TTC CTA ATT G. Products: mutant allele 754 bp, wild-type allele 630 bp.