Cathepsin B and D deficiency in the mouse pancreas induces impaired autophagy and chronic pancreatitis

The major lysosomal proteases, Cathepsin B (CTSB), Cathepsin D (CTSD) and Cathepsin L (CTSL), are implicated in autophagic activity. To investigate the role of each cathepsin in the exocrine pancreas, we generated mice in which the pancreas was specifically deficient in Ctsb, Ctsd and Ctsl. Each of these gene knockout (KO) and Ctsb;Ctsl and Ctsd;Ctsl double-knockout (DKO) mice were almost normal. However, we found cytoplasmic degeneration in the pancreatic acinar cells of Ctsb;Ctsd DKO mice, similar to autophagy related 5 (Atg5) KO mice. LC3 and p62 (autophagy markers) showed remarkable accumulation and the numbers of autophagosomes and autolysosomes were increased in the pancreatic acinar cells of Ctsb;Ctsd DKO mice. Moreover, these Ctsb;Ctsd DKO mice also developed chronic pancreatitis (CP). Thus, we conclude that both Ctsb and Ctsd deficiency caused impaired autophagy in the pancreatic acinar cells, and induced CP in mice.

Autophagy is a system of intracellular degradation that involves lysosomal enzymes. The role of autophagy is balancing sources of energy at critical times in development and in response to nutrient stress 1 . Moreover, autophagy plays a critical role in clearing misfolded or aggregated proteins, removing damaged organelles, such as mitochondria, endoplasmic reticulum and peroxisomes 1 . Autophagy exerts devastating effects in pancreatic acinar cells by the activation of trypsinogen to trypsin in the early stage of acute pancreatitis 2 . However, basal autophagy maintains pancreatic acinar cell homeostasis and protein synthesis 3 and impaired autophagy induces chronic pancreatitis (CP) [4][5][6][7] .
Cathepsin B (CTSB) and L (CTSL) are cysteine proteases, and Cathepsin D (CTSD) is an aspartic protease 8 . These cathepsins are major lysosomal proteases and are widely expressed in endosomes and lysosomes. Mice lacking Ctsd in all tissues die one month after birth due to intestinal necrosis 9 . Mice lacking Ctsb or Ctsl in all tissues grow up almost normally [10][11][12] , however, mice lacking both Ctsb and Ctsl in all tissues die 2 weeks after birth due to the massive apoptosis of central neurons 13 . Thus, the organ-specific roles of these cathepsins are different. Furthermore, Ctsb and Ctsl were reported to be involved in the processing of CtsD 14,15 , and cathepsins might interact with each other in different ways depending on tissues 16,17 .
On the other hand, Ctsb and Ctsl play a role in intrapancreatic trypsinogen activation and the onset of acute pancreatitis [18][19][20] . Ctsd was reported to regulate Ctsb activation during experimental pancreatitis 17 , and was implicated in Ctsb and Ctsl degradation, but not in autophagic activity in the pancreas 21 . Thus, it is not clear which of the cathepsins plays an important role in autophagy of the pancreas.
The expression of autophagy markers was normal in the pancreas of Ctsb ΔPan , Ctsd ΔPan and Ctsl ΔPan mice. In order to investigate the autophagic activity in the pancreases of Ctsb ΔPan , Ctsd ΔPan or Ctsl ΔPan mice, we firstly performed histopathological examinations of the pancreases of Ctsb ΔPan , Ctsd ΔPan and Ctsl ΔPan mice. These examinations revealed no abnormalities (Fig. 1Ba-d).
We then investigated the expression of p62/sequestosome1 (p62), a selective substrate of autophagy, and microtubule-associated protein 1 light chain 3 (LC3), a marker of autophagic vacuole formation, in order to analyze the autophagy function in the pancreases of Ctsb ΔPan , Ctsd ΔPan and Ctsl ΔPan mice. As shown in Fig. 1C, no significant difference in the expression of p62 in the pancreases was noted among wild-type, Ctsb ΔPan , Ctsd ΔPan and Ctsl ΔPan mice (Fig. 1C). LC3 has two forms. LC3-I is localized in the cytoplasm and is converted into LC3-II. LC3-II is associated with the membranes of autophagosomes in the phosphatidylethanolamine conjugated form, and the amount of LC3-II is correlated with the extent of autophagosome formation. In the pancreases of wildtype, Ctsb ΔPan , Ctsd ΔPan and Ctsl ΔPan mice, there was no apparent difference in the ratio of LC3-II/Actin (Fig. 1E).
We then investigated the expression of p62 and LC3 in order to analyze the autophagy function in the pancreases of Ctsb ΔPan ;Ctsd ΔPan , Ctsb ΔPan ;Ctsl ΔPan or Ctsd ΔPan ;Ctsl ΔPan mice. As shown in Fig. 2C, there were no significant differences in the expressions of p62 and LC3 among wild-type, Ctsb ΔPan ;Ctsl ΔPan and Ctsd ΔPan ;Ctsl ΔPan mice. On the other hand, Ctsb ΔPan ;Ctsd ΔPan mice showed a marked increase in the autophagy substrates p62 and the membrane-bound form of LC3 (Fig. 2C). Furthermore, we performed immunofluorescent staining of the pancreases of wild-type and Ctsb ΔPan ;Ctsd ΔPan mice, which demonstrated the accumulation of p62 and LC3 in the acinar cells of Ctsb ΔPan ;Ctsd ΔPan mice ( Fig. 2Ea-l).
Autolysosomes accumulated in pancreatic acinar cells of Ctsb ΔPan ;Ctsd ΔPan mice, while autophagosomes were also present in the cells but fewer in number. In order to clarify which phase of the autophagic pathway was disturbed in the pancreas of Ctsb ΔPan ;Ctsd ΔPan mouse pancreas, in the view of the morphology, we observed the pancreas using a transmission electron microscope (TEM). A large number of granular structures were observed in the pancreatic acinar cells of Ctsb ΔPan ;Ctsd ΔPan mice compared with wildtype mice (Fig. 3A,B,E,F). In the cytoplasm of the acinar cells of Ctsb ΔPan ;Ctsd ΔPan mice, most granular structures that possessed cytoplasmic organelles with degraded or undegraded materials, in particular the granular endoplasmic reticulum with a large amount of synthetized materials, were autolysosome-like structures (Fig. 3A,B). In addition to such autolysosome-like structures, the Ctsb ΔPan ;Ctsd ΔPan acinar cells had multi-membranous vacuolar structures that corresponded to the autophagosomes often seen in cathepsin D-defective neurons 24 (Fig. 3A www.nature.com/scientificreports/ acinar cells often contained large granules of cellular debris in which zymogen granules and nuclear structures were detected (Fig. 3C). In some cases, inclusion bodies which contained a nucleus with condensed chromatin were detected in acinar cells (Fig. 3D). In contrast, wild-type acinar cells possessed lysosome/autolysosome-like structures in the cytoplasm, but no clear-cut autophagosome-like structures were detected in the cells, in contrast to Ctsb ΔPan ;Ctsd ΔPan acinar cells (Fig. 3E,F).

Chronic pancreatitis was induced in Ctsb ΔPan
;Ctsd ΔPan mice. We compared the weight at one, four and eight months old between wild-type mice and Ctsb ΔPan ;Ctsd ΔPan mice. No significant difference was noted www.nature.com/scientificreports/ www.nature.com/scientificreports/ www.nature.com/scientificreports/ between the weights at one and four months old, but the weight at eight months old was significantly lower in Ctsb ΔPan ;Ctsd ΔPan mice than in wild-type mice (Fig. 4A). On a histopathologic examination of the pancreas in four-month-old Ctsb ΔPan ;Ctsd ΔPan mice, numerous vacuoles were observed in the acinar cells, in comparison to wild-type mice (Fig. 4Ba, b, d). The loss of the normal architecture and replacement of acinar cells with fat was observed in eight-month-old Ctsb ΔPan ;Ctsd ΔPan mice (Fig. 4Bc). Pancreatic fibrosis in the pancreases of four-month-old and eight-month-old Ctsb ΔPan ;Ctsd ΔPan mice was confirmed by Azan-Mallory staining (Fig. 4Be, f, Ci). Fibrosis and replacement of acinar cells with fat worsened over time in the pancreas of eight-month-old Ctsb ΔPan ;Ctsd ΔPan mice compared with the pancreas of four-month-old Ctsb ΔPan ;Ctsd ΔPan mice. Immunostaining of the pancreases in four-month-old Ctsb ΔPan ;Ctsd ΔPan mice revealed a large number of inflammatory cells, including monocytes (CD11b-positive cells), macrophages (F4/80-positive cells) and plasma cells (CD138-positive cells) infiltrated the pancreas (Fig. 4Ce-h, j-l). No monocyte, macrophage nor plasma cell was detected in the pancreases of four-month-old wild-type mice (Fig. 4Ca-d). Immunofluorescent staining of the pancreas in four-month-old mice showed chronic inflammatory cell infiltration and progressive fibrosis (Fig. 4D).

Discussion
The purpose of this study is to clarify the role of cathepsin B, D and L in autophagy of the pancreas. The pancreas of Ctsb ΔPan ;Ctsd ΔPan mice showed impaired autophagy. On the other hand, autophagy was not impaired in the pancreases of Ctsb ΔPan and Ctsd ΔPan mice. These results indicate that Ctsb and Ctsd play an important role and act synergistically in the autophagy of the pancreas.
Ctsd plays a critical role in the autophagy of the intestine 9 and the autophagy of central nerves requires Ctsb or Ctsl 13 . This study proved that the autophagy of pancreatic acinar cells in mice required both Ctsb and Ctsd. In this study, the autophagy in the pancreas was not impaired in Ctsb ΔPan , Ctsd ΔPan or Ctsl ΔPan mice, as reported previously [10][11][12]21 . The Ctsb and Ctsl expression was increased in the pancreas of Ctsd ΔPan mice. The same result was reported previously 21 . In this study, the expression of LC3-II was not increased in the pancreas of Ctsd ΔPan mice. A previous report 21 evaluated the ratio of LC3-II to LC3-I. LC3-II should be calculated using the ratio of a housekeeping protein, not LC3-I, as LC3-I tends to be less sensitive to detection by certain anti-LC3 antibodies 25 . This is the reason why we calculated the ratio of LC3-II to actin. In addition, the serine protease inhibitor Kazal type 3 (Spink3), which is a trypsin specific inhibitor in the pancreas, -cre knock-in mice were used in the previous report 21 , while pancreas transcription factor 1 subunit alpha (Ptf1a) -cre knock-in mice (Ptf1a cre/+ ) were used in this study. Spink3 expression was reported in not only the pancreas, but also in the intestine, kidney and epididymis, among other sites 26 . Ptf1a expression would be more specific to the pancreas than Spink3, so we used Ptf1a cre/+ mice in this study.
The expression of Ctsb in Ctsb ΔPan ;Ctsd ΔPan mice did not disappeared. However, the expression ratio of pro-Ctsb/Ctsb heavy chain in Ctsb ΔPan ;Ctsd ΔPan mice was different from in wild-type mice. Ctsd in pancreatic acinar cells was shown to be involved in Ctsb and Ctsl degradation 21 and regulated Ctsb activation under conditions of experimental pancreatitis 17 . Ctsd was shown to be accumulated in the brain of Ctsb;Ctsl DKO mice in all cells 13 . These results indicate that Ctsb, Ctsd and Ctsl regulate each other in vivo. The Ctsb ΔPan ;Ctsd ΔPan mice showed a marked increase in the autophagy substrates p62 and LC3. This result indicates that impaired autophagy was induced in the Ctsb ΔPan ;Ctsd ΔPan mouse pancreas. Because Ctsb and Ctsd are lysosomal enzymes, autophagy in the Ctsb ΔPan ;Ctsd ΔPan mouse pancreas could be disturbed in the degradation phase in autolysosomes. The accumulation of autolysosomes with a few autophagosomes in the pancreatic acinar cells of Ctsb ΔPan ;Ctsd ΔPan mice was consistent with this expectation.
In the histopathologic examination, the Ctsb ΔPan ;Ctsd ΔPan mouse pancreas showed CP, which progressed with time. At least four animal models of CP with impaired autophagy have been previously reported: mice with pancreas-specific disruption of Atg5 that developed a form of CP 4 , mice with pancreas-specific ablation of IκB kinase α 5 , mice lacking Spink3 with a mosaic pattern of SPINK1 expression 6 , and LAMP-2-deficient mice 7 . The histological findings of CP, such as inflammation, acinar-to-ductal metaplasia and acinar-cell hypertrophy, have been confirmed in all these models. Additional experiments have been performed, such as qRT-PCR analyses of pancreatic RNA from fibrogenic markers and cytokine and chemokine genes and immune cell markers, assessments of serum levels of pancreatitis markers, evaluations of changes in the body and pancreas weight and immunostaining of the pancreas. On the other hand, impaired autophagy was evidenced by accumulation of p62 in WB, enlarged autophagic vacuoles in HE staining and TEM of pancreatic acinar cells. Immunostaining of the pancreas was also performed as an additional experiment. Ctsb ΔPan ;Ctsd ΔPan mice may be another animal model of CP with impaired autophagy.
In 1896, Chiari reported that pancreatitis was the result of pancreatic autodigestion 27,28 . In 1959, Greenbaum reported Ctsb activated trypsinogen to trypsin in vitro 29 . Steer suggested that Ctsb might be responsible for the intracellular activation of digestive enzymes [30][31][32] . It became clear that Ctsb inhibition reduced the severity of pancreatitis 33 , the release of Ctsb in cytosol caused cell death in acute pancreatitis 34 , and Ctsb activity initiated experimental pancreatitis 35 . Ctsb, as described above, is known to be deeply implicated in pancreatitis. The present study also indicated that Ctsb played an important role of autophagy in pancreas. Ctsd has many functions as an aspartic protease [36][37][38] , that interacts with other important molecules and influences cell signals [39][40][41] . Ctsd is implicated in apoptosis and cancer, but not in the severity of pancreatitis or autophagy activity by itself 21 . From the above results, both Ctsb and Ctsd play an important synergistic role in the autophagy of the mouse pancreas. www.nature.com/scientificreports/ www.nature.com/scientificreports/ In conclusion, both Ctsb and Ctsd deficiency caused impaired autophagy in pancreatic acinar cells and induced CP in mice. A future challenge is to clarify the mechanism underlying the interaction of Ctsb and Ctsd with regard to autophagy.

Materials and methods
Animal protocol and experimental design. Mice were kept under specific-pathogen-free conditions with ad libitum access to food and water in a 12 h (h) light/dark cycle. C57BL/6 N mice were purchased from CREA Japan. All animal experiments were performed with the approval of the Hyogo College of Medicine Institutional Animal Care and Use Committee , and the ARRIVE guideline. All methods were carried out in accordance with the relevant guidelines of the Hyogo College of Medicine and the ARRIVE guideline, including any relevant details. Histological and immunohistochemical analyses. For the histological analyses, pancreatic tissue was fixed overnight in 15% formalin, embedded in paraffin, sectioned, and subjected to HE and Azan-Mallory staining. Immunohistochemistry was performed using the antibodies listed above.
Western blotting. Pancreases were disrupted with TissueLyser LT (QIAGEN) and homogenized in RIPA buffer. The homogenates, which included 80 μg proteins, were subjected to SDS-PAGE (ATTO) and transferred onto Immobilon polyvinylidene difluoride membranes (Millipore). Blots were incubated with 5% nonfat dry milk in phosphate-buffered saline with 0.1% Tween 20 at room temperature for 1 h to block nonspecific binding, overnight at 4 °C with primary antibodies in blocking buffer, and finally with horseradish peroxidase conjugated secondary antibodies in blocking buffer at room temperature for 1 h. The membranes were developed with Chemi-Lumi One L and super (Nacalai Tesque), and analyzed in a WSE-6100H LuminoGraph1 (ATTO). Akt and actin were used as loading controls.
Transmission electron microscopy. Anesthetized mice were fixed with 2% glutaraldehyde and 2% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4. Pancreatic tissues were extracted after fixation. Slices of the fixed tissues were postfixed with 2% OsO4, dehydrated in ethanol and embedded in Epok 812 (Okenshoji Co.). Ultrathin sections were cut with an ultramicrotome (ultracut N or UC6: Leica). These sections were stained with uranyl acetate and lead citrate and examined on a Hitachi HT7700 or JEOL JEM-1230 electron microscope.
Reverse transcriptase RT-PCR. Total RNA was isolated using an RNeasy Plus Universal Mini Kit (Qiagen). cDNA was synthesized using a ReverTra Ace qPCR RT Kit (Toyobo). For the detection of Il6 mRNA, Tnfα and Il1β, a Thermal Cycler Dice Real Time PCR System III was used with SYBR Premix Ex Taq (Takara). www.nature.com/scientificreports/