The CEL-HYB1 Hybrid Allele Promotes Digestive Enzyme Misfolding and Pancreatitis in Mice

Abstract

Background & aims: A hybrid allele that originated from homologous recombination between CEL and its pseudogene (CELP), CEL-HYB1 increases the risk of chronic pancreatitis (CP). Although suggested to cause digestive enzyme misfolding, definitive in vivo evidence for this postulate has been lacking.

Methods: CRISPR-Cas9 was used to generate humanized mice harboring the CEL-HYB1 allele on a C57BL/6J background. Humanized CEL mice and C57BL/6J mice were used as controls. Pancreata were collected and analyzed by histology, immunohistochemistry, immunoblotting, and transcriptomics. Isolated pancreatic acini were cultured in vitro to measure the secretion and aggregation of CEL-HYB1 protein. Mice were given caerulein injections to induce acute pancreatitis (AP) and CP.

Results: Pancreata from mice expressing CEL-HYB1 developed pathological features characteristic of focal pancreatitis that included acinar atrophy and vacuolization, inflammatory infiltrates, and fibrosis in a time-dependent manner. CEL-HYB1 expression in pancreatic acini led to decreased secretion and increased intracellular aggregation and triggered endoplasmic reticulum stress compared with CEL. The autophagy levels of pancreata from mice expressing CEL-HYB1 changed at different developmental stages; some aged CEL-HYB1 mice exhibited an accumulation of large autophagic vesicles and impaired autophagy in acinar cells. Administration of caerulein increased the severity of AP/CP in mice expressing CEL-HYB1 compared with control mice, accompanied by higher levels of endoplasmic reticulum stress.

Conclusions: Expression of a humanized form of CEL-HYB1 in mice promotes endoplasmic reticulum stress and pancreatitis through a misfolding-dependent pathway. Impaired autophagy appears to be involved in the pancreatic injury in aged CEL-HYB1 mice. These mice have the potential to be used as a model to identify therapeutic targets for CP.

Keywords: Carboxyl Ester Lipase; Chronic Pancreatitis; Genetic Variants; Protein Misfolding.

Graphical abstract

Figure 1

Expression of tagged and untagged wild-type hCEL and CEL-HYB1 in HEK293T cells. HEK293T cells were transiently transfected with plasmids encoding wild-type hCEL and CEL-HYB1 with or without a 3×Flag tag (+/−). (A) Cell medium and soluble (lysate) and insoluble (pellet) fractions were harvested after 48 hours and were analyzed by Western blotting. Each blot is representative of 4 independent experiments. Empty vector (EV) (negative control). β-actin was measured as a loading control. (B) Quantification of Western blot band intensities adjusted to the β-actin levels. (C) Expression of ER stress markers HSPA5 and sXBP1 were assessed by Western blotting of the soluble (lysate) fractions. (D and E) Quantification of marker band intensities after adjustment to the β-actin levels and normalization to the expression in EV-transfected cells. Throughout the Figure, data are presented as mean values ± standard deviation (SD) (n = 4); bars with different letters represent significant differences in means by 1-way analysis of variance with Sidak’s multiple comparisons.

Figure 2

Generation of humanized carboxyl ester lipase (CEL) mice and expression analysis. (A) Targeting sequences (containing Flag tag) were designed for developing humanized CEL or CEL-HYB1 mice and (B) inserted into exon 2 of mouse Cel separately by CRISPR/Cas9 gene editing. (C) The signal peptide of CEL protein expressed by humanized mice is derived from mouse Cel. (D) Expression of CEL protein in the soluble fraction from mice aged 12 weeks was detected by Western blot analysis using anti-CEL (upper) and anti-Flag antibodies (middle). β-actin was measured as a loading control (down). (E) Immunohistochemical analysis of CEL protein expression in the pancreata of C57BL/6J, hCEL-HYB1, and hCEL mice. Scale bar = 100 μm. (F) Pancreatic mRNA expression of mouse Cel, hCEL-HYB1, and hCEL using a primer pair specific for hCEL-HYB1 and hCEL; Gene expression levels are expressed as fold changes relative to Gapdh mRNA levels. (G) mRNA expression of hCEL-HYB1 and hCEL relative to mouse Cel. Mean ± standard deviation (SD). Significance (exact P values as indicated) was determined by 1-way analysis of variance with Sidak’s multiple comparisons. Homozygous mice of both hCEL-HYB1 and hCEL were used for Western blot, immunohistochemistry, and quantitative PCR analysis. (H) Western blot analysis of CEL protein expression in the pancreas of homozygous (hCEL-HYB1+/+) and heterozygous (hCEL-HYB1+/−) mice. β-actin was measured as a loading control.

Figure 3

Histological analysis of C57BL/6J, hCEL-HYB1, and hCEL mice. (A) Body weight of each mouse strain at 1 month, 3 months, 6 months, and 12 months of age. (B) The relative pancreas weights of each mouse strain, expressed as a percentage of body weight, at 12 months of age. Mean ± standard deviation (SD). ∗P < .05; ∗∗P < .01; 1-way analysis of variance with Sidak’s multiple comparisons. (C) Representative hematoxylin and eosin (H&E)-stained tissue sections of pancreas isolated from homozygous hCEL-HYB1, hCEL, or wild-type C57BL/6J mice of different ages. Representative H&E images from 4-month-old hCEL-HYB1 mice showing mild vacuolization of the acinar cells (top). Pancreatic section from an 8-month-old hCEL-HYB1 mouse showing focal immune cell infiltration and loss of adjacent acinar cells (middle). More severe inflammatory infiltration, fibrotic reaction, and fatty changes (red dotted box) in pancreatic tissue from a 14-month-old hCEL-HYB1 mouse (bottom). Scale bar = 100 μm. (D) Representative H&E-stained tissue section of pancreas isolated from a heterozygous hCEL-HYB1 mouse at 14 months of age showing inflammatory infiltration (black arrow), fatty changes (green arrow), and acinar atrophy (red arrow). Scale bar = 1 mm. (E) Masson’s trichrome staining showing fibrotic changes (blue color) resulting from the destruction of acini. Scale bar = 100 μm. (F and G) Immunohistochemistry for Cd45 and F4/80 demonstrates focal areas of immune cell infiltration in the pancreata of hCEL-HYB1 mice. Scale bar = 100 μm. (H and I) Quantification under the light microscope of Cd45-positive and F4/80-positive cells for C57BL/6J, hCEL-HYB1, and hCEL mice aged 12 months. Mean ± SD (n = 8). Significance (exact P values indicated) was determined by 1-way analysis of variance with Sidak’s multiple comparisons.

Figure 4

Apoptosis, focal injury, vacuolization, and eosinophilic inclusion bodies of pancreatic acinar cells in hCEL-HYB1 mice. (A) Immunohistochemistry for the apoptosis marker, cleaved caspase-3, as well as (B) TUNEL assay, used to detect DNA breaks formed during apoptosis, performed on pancreas sections from C57BL/6J, hCEL-HYB1 and hCEL mice at 10 months of age. Red arrows mark representative cleaved caspase-3-positive cells. Scale bar = 100 μm. (C) Quantitative analysis of TUNEL-positive cell staining in pancreatic sections showed that 1-year-old hCEL-HYB1 mice developed higher levels of apoptosis than controls. Mean ± standard deviation (SD). Significance (exact P values as indicated) was determined by 1-way analysis of variance with Sidak’s multiple comparisons. (D) Pancreas from a 10-month-old hCEL-HYB1 mouse exhibited focal destruction of acinar cells, and immunohistochemical analysis for Hmgb1, a characteristic marker of cellular damage. Red arrows mark negative staining of Hmgb1 in the nucleus, and black arrows mark positive staining of Hmgb1 in the cytoplasm. Scale bar = 100 μm. (E) Vacuolization and eosinophilic inclusion bodies (black arrows) of acinar cells in the pancreas of a 12-month-old hCEL-HYB1 mouse. Immunohistochemistry for hCEL-HYB1 protein with anti-Flag antibody (right). Scale bar = 100 μm.

Figure 5

Proteotoxic misfolding and ER stress in hCEL-HYB1 mice. (A) Protein content of hCEL and hCEL-HYB1 in intracellular fraction, medium, and soluble and insoluble fractions from isolated acinar cells were analyzed by Western blotting using an anti-Flag antibody. Data are representative of 4 independent experiments. (B) Estimation of CEL content by densitometry of immunoblots (n = 4). For each subtype, CEL content in the intracellular fraction was arbitrarily set to 1.0. Mean ± standard deviation (SD). ∗P < .05; ∗∗P < .01; ∗∗∗∗P < .0001; ns, P > .05; 2-tailed unpaired Student t test. (C) Immunoblots of sXbp1 (soluble), Ddit3 (soluble), Hspa5 (soluble and insoluble) in pancreatic tissue homogenates of hCEL-HYB1, hCEL, and C57BL/6J mice aged 5 months. (D) Quantification of bands after adjustment to the β-actin levels. Mean ± SD (n = 6). Bars with different letters represent significant differences in means by 1-way analysis of variance with Sidak’s multiple comparisons. (E-I) Messenger RNA expression of Ddit3, Hspa5, Hspa1a, Hsp90aa1, and Hsp90ab1 was measured by quantitative PCR. Results were expressed as fold change relative to the C57BL/6J results. Mean ± SD (n = 8). Significance (exact P values indicated) was determined by 1-way analysis of variance with Sidak’s multiple comparisons. (J) Pancreatic trypsin activity in C57BL/6J, hCEL-HYB1, and hCEL mice at 5 months of age showing no significant difference. Data shown are mean ± SD; 1-way analysis of variance test.

Figure 6

Autophagy in hCEL-HYB1 mice. (A) Representative anti-LC3 immunofluorescence staining in pancreatic sections from mice aged 8 weeks and 24 weeks. The images display LC3 signals in green, and cell nuclei in blue. Scale bar = 100 μm. (B) Expression levels of p62, beclin1, and LC3 in pancreatic homogenates from C57BL/6J, hCEL-HYB1, and hCEL mice aged 8 weeks and 24 weeks were detected by Western blot (upper panel). Bottom, quantification of bands after adjustment to the β-actin. Mean ± standard deviation (SD) (n = 6). Bars with different letters represent significant differences in means by 1-way analysis of variance with Sidak’s multiple comparisons. (C) Pancreatic section from a 40-week-old hCEL-HYB1 mouse showing cytoplasmic vacuolization of acinar cells (upper). LC3 immunofluorescence staining in pancreatic sections with severely vacuolated acinar cells (down). Scale bar = 100 μm. (D and E) Co-localization of LC3-positive vacuoles with hCEL-HYB1 (Flag) or Lamp1 in pancreatic sections from a 40-week-old hCEL-HYB1 mouse with severe vacuolation of acinar cells. Scale bar = 40 μm. (F) Expression levels of p62, beclin1, and LC3 in pancreatic homogenates from hCEL-HYB1 mice with severe cytoplasmic vacuolation of acinar cells were detected by Western blot analysis (left panel). Right, quantification of bands after adjustment to the β-actin. Mean ± SD (n = 4). Bars with different letters represent significant differences in means by 1-way analysis of variance with Sidak’s multiple comparisons. (G-I) Immunohistochemical analysis for p62, cleaved caspase-3, and Hmgb1 in early vacuolated acinar cells. Red dotted line outlines the area of severely vacuolated acinar cells showing negative cleaved caspase-3 staining (middle). Green dotted line outlines the vacuolated acinar cells without obvious nuclear-cytoplasmic translocation of Hmgb1 (right). Scale bar = 40 μm.

Figure 7

Ultrastructural changes in hCEL-HYB1 mice. (A) Representative transmission electron micrographs of pancreas from C57BL/6J, hCEL-HYB1, and hCEL mice aged 5 months. hCEL-HYB1 mice show slightly dilated endoplasmic reticula and vacuolation (black arrow). Scale bar = 4 μm. (B) Degradative vacuoles in an acinar cell from an hCEL-HYB1 mouse aged 6 months. Scale bar = 0.5 μm. Pancreatic tissue sections from 12-month-old hCEL-HYB1 mice showing protein aggregates in cytoplasm (Scale bar = 1 μm) (C); large autophagy vesicles with undegraded high-density contents (Scale bar = 2 μm) (D); as well as evident dilated endoplasmic reticula in some acinar cells (Scale bar = 4 μm) (E).

Figure 8

Caerulein-induced acute pancreatitis in C57BL/6J, hCEL-HYB1, and hCEL mice. (A) For induction of AP, hCEL-HYB1 and control mice aged 6 months old were given 12-hourly injections of caerulein (50 μg/kg/h). (B) Representative images of hematoxylin and eosin (H&E)-stained tissue sections isolated from C57BL/6J, hCEL-HYB1, and hCEL mice. Scale bar = 100 μm. (C) Immunohistochemical analysis for apoptotic cells with anti-cleaved caspase-3 antibody. Scale bar = 100 μm. (D) Semi-quantitative histology score examination of H&E-stained sections from 3 mouse strains. (E) Serum amylase activity. (F-H) Messenger RNA expression of ER stress markers Atf4, Hspa5, and Ddit3 was measured by quantitative PCR. (I-K) Messenger RNA expression of Il6, Tnf, and Bcl2. Throughout the Figure, data are presented as mean values ± standard deviation (SD); bars with different letters represent significant differences in means by 1-way analysis of variance with Sidak’s multiple comparisons. NS, normal saline; Cer, caerulein.

Figure 9

Caerulein-induced chronic pancreatitis in C57BL/6J, hCEL-HYB1, and hCEL mice. (A) Flow diagram of caerulein-induced CP: 6 daily caerulein injections (50 ug/kg body weight), given 1 hour apart, on days 1, 3, and 5 of each week, for 4 weeks. (B) Representative images of hematoxylin and eosin (H&E)- and Sirius red-stained tissue sections isolated from C57BL/6J, hCEL-HYB1, and hCEL mice. Scale bar = 100 μm. (C) Relative pancreas weights (shown as a percentage of body weight). (D) Histologic score and quantitative analysis. (E-F) Messenger RNA expression of Acta2 and Hspa5 was measured by quantitative PCR. Throughout the Figure, data are presented as mean values ± standard deviation (SD); bars with different letters represent significant differences in means by 1-way analysis of variance with Sidak’s multiple comparisons. NS, normal saline; Cer, caerulein.

Figure 10

Recombined hCEL/hCEL-HYB1 targeting sequence and genotyping of mouse strains. (A and B) The 5′ homology arm and the 3′ homology arm are underlined. The exons of mouse Cel are highlighted in green. The 5′ homology arm contains exon 1 plus the proximal portion of exon 2 of mouse Cel (including the intervening intron). The 3′ homology arm contains the remainder of exon 2 plus exons 3–7 of mouse Cel (including the introns). hCEL/hCEL-HYB1 exons 2–11 are marked in yellow, the VNTR sequences are indicated in italics, the 3×Flag is in red, and the 3′UTR-polyA is in blue. (C) Primer design for detecting wild-type and novel knock-in strains. Primers used to genotype mouse Cel: 5′-gcggtgggaaagtgcggggatagt-3′ (P1) and 5′-cttgccagccagggtgacg-3′ (P2). Primers used to genotype hCEL-HYB1 or hCEL: 5′-gctagagcttggcgtaatcat-3′ (P3) and 5′-cagtcttcttgcccataggtgt-3′ (P4). (D) A representative agarose gel image of the genotyping. The mouse Cel amplicon size was 697 bp, whereas both hCEL-HYB1 and hCEL yielded 402 bp PCR products.