Immunoproteasome impairment via β5i/LMP7‐deletion leads to sustained pancreatic injury from experimental pancreatitis

Abstract Uncovering potential new targets involved in pancreatitis may permit the development of new therapies and improvement of patient's outcome. Acute pancreatitis is a primarily sterile disease characterized by a severe systemic inflammatory response associated with extensive necrosis and a mortality rate of up to 24%. Considering that one of the reported disease mechanisms comprises the endoplasmic reticulum (ER) stress response and that the immunoproteasome is a key regulator to prevent proteotoxic stress in an inflammatory context, we investigated its role in acute pancreatitis. In this study, we demonstrate that immunoproteasome deficiency by deletion of the β5i/LMP7‐subunit leads to persistent pancreatic damage. Interestingly, immunoproteasome‐deficient mice unveil increased activity of pancreatic enzymes in the acute disease phase as well as higher secretion of Interleukin‐6 and transcript expression of the Interleukin IL‐1β, IFN‐β cytokines and the CXCL‐10 chemokine. Cell death was increased in immunoproteasome‐deficient mice, which appears to be due to the increased accumulation of ubiquitin‐protein conjugates and prolonged unfolded protein response. Accordingly, our findings suggest that the immunoproteasome plays a protective role in acute pancreatitis via its role in the clearance of damaged proteins and the balance of ER stress responses in pancreatic acini and in macrophages cytokine production.

by accumulation of misfolded or unfolded proteins, which may arise during synthesis, folding and secretion of secretory and cell-surface proteins. Therefore, elucidating the molecular mechanisms of protein homeostasis may identify potential new targets of acinar cell injury.
ER stress is counterbalanced by the unfolded protein response (UPR) machinery, which has a dedicated role in protein quality control and homeostasis, interacting with degradation pathways such as autophagy and the ubiquitin proteasome system (UPS). 17 Impairment of UPR was found to be one cause of genetic varieties of pancreatic disorders. 18,19 The UPS plays an important role in intracellular protein degradation and turnover in eukaryotes via a multi-enzymatic machinery, entailing target protein ubiquitination and subsequent proteolysis by the 26S proteasome. 20 The immunoproteasome is an isoform of the constitutive proteasome (26S proteasome), which is assembled upon a triggering inflammatory response by Toll-like or cytokine receptor signalling, type I or type II interferons (IFNs), and TNFα releases. It arises from the replacement of the β-catalytic subunits β1, β2 and β5 in the 20S standard proteasome core by β1i (large multifunctional peptidase 2, LMP2 encoded by Psmb9), β2i (multicatalytic endopeptidase complex-like-1, MECL-1 encoded by Psmb10) and β5i (LMP7 encoded by Psmb8), respectively, during de novo proteasome formation. 21,22 Over the years, the immunoproteasome has been recognized as an important player in shaping innate and adaptive immune responses by degradation of inflammatory mediators and improved MHC class I antigen presentation. 23 In addition to its roles in innate and adaptive immune cells, the immunoproteasome has also been linked to non-immune functions, namely cell differentiation and protein homeostasis, including intracellular protein clearance 24 and subsequently, may thus control of the ER stress levels. Likewise, a growing body of evidence has unveiled a consistent interdependence between ER stress imbalance and the immune system. 25,26 In view of the immunoproteasome role for protein clearance and ER stress pathways, and the association of human mutations with autoinflammatory diseases, 27 our aim in this work was to investigate the role of the β5i/LMP7 subunit in acute pancreatitis.

| Animals
Experiments were performed using male and female β5i/LMP7 +/+ (littermate controls) and β5i/LMP7 −/− mice C57BL/6J strain background, generated as previously reported, 28 weighing 20 to 25g at about 2-3 months of age. Both genotypes were bred and maintained at the Greifswald University Animal Care Facility. Food and water were provided ad libitum to the animals, which were kept in a controlled environment with a constant 12:12-hours light-dark cycle.

| Caerulein-induced acute pancreatitis model
Acute pancreatitis was induced by eight hourly intraperitoneal injections of caerulein (50 µg/kg bodyweight; cat: C9026-Sigma, Merck), as previously described. 29 The experimental protocol was approved by the institutional Animal Care and Use Committee.
Three experimental groups were designed: control animals (0 hour; no caerulein injections), 8 hours corresponding to the acute phase of the disease and 24 hours to the recovery phase. Five independent experiments were performed, and no difference regarding sex was statistically significant. Serum and tissues were collected, processed and adequately stored to the experimental procedures, as described ahead.

| Biochemical assays
Quantification of serum activities of amylase and lipase enzymes were assayed by photometric assays kits from Roche Hitachi. Lactate dehydrogenase (LDH) was measured in serum using a specific kit assay from Sigma (cat: MAK066, Merck). Pancreatic enzymatic activities, such as trypsin and chymotrypsin, were determined in tissue homogenates as previously reported. 30 Proteasome Chymotrypsin-like activity assay was also performed in pancreas homogenates, as described in supplementary method. Myeloperoxidase (MPO) activity measurement in lung was performed as previously described. 11 All activities from tissue homogenates were suitably normalized by protein concentration using Pierce BCA protein assay (cat: 23227, Thermo Fisher Scientific).

| Measurement of pro-inflammatory cytokines
The IL-6 pro-inflammatory cytokine was measured in serum and supernatant samples by fluorescence activated cell sorter (FACS) analysis using the CBA mouse inflammation kit, according the manufacturer's instructions (cat: 552364; Becton Dickinson). IL-1β cytokine was measured in supernatant from macrophage culture using Peprotech kit assay (cat: 900-M47).
A list of all used antibodies can be found in supplementary methods. Quantification of the CD68 and Ly6G immune staining was estimated manually from 5 random pictures in a magnification of 200-fold from each animal. Total number of cells, through WEKA Segmentation in Image J software, was used to obtain the percentage of positive cells. The same design was applied to calculate apoptosis ratio.

| Isolation of acini
Pancreas were carefully removed from wild-type C57BL/6J, β5i/ LMP7 +/+ (littermate controls) and β5i/LMP7 −/− mice and digested with 1 mg collagenase (Collagenase from Clostridium histolyticum; EC.3.4.24.3; SERVA Electrophoresis), under sterile conditions, as initially described. 32 Cell suspension was resuspended in fresh medium without collagenase, and after 30 minutes of resting in water bath at 37℃, three different approaches were addressed: 1) protease activation was examined in living cells upon an enzymatic kinetic reaction up to 60 minutes with 1 µM Cholecystokinin (CCK, fragment 26-33; C-2175 Sigma) stimulation whereas necrosis was measured by propidium iodide exclusion 33 ; 2) acini were directly platted in 12 well plates and incubated with 1 µM CCK for 6 hours or 8 hours and 24 hours in a humidified atmosphere of 5% CO 2 at 37ºC, for subsequent total RNA or protein extraction; 3) or challenged with the 1 µM of CCK for 30 minutes prior addition to the macrophage culture, as described below.

| Co-culture of acini and macrophages
Femur of wild-type C57BL/6J, β5i/LMP7 +/+ (littermate controls) and β5i/LMP7 −/− mice were collected under sterile conditions. Bone marrow stem cells were differentiated into macrophages (BMDM), as previously established. 34 At the day of the experiment, BMDM culture received fresh medium, before being exposed to isolated acini for 6 hours or 9 hours. Supernatant was collected for cytokine measurements, and the cells were washed with PBS to withdraw non-adherent cells. Subsequently, cells were harvested with 5 mM EDTA in PBS on ice and resulting pellet stored at −80ºC for posterior total RNA extraction.

| Transcript expression by quantitative Realtime PCR
All reagents were purchased from Thermo Fisher Scientific, unless otherwise specified. Pancreas, acini and macrophages were collected and total RNA was extracted with Trizol reagent (cat: 15596026). RT-qPCR analyses are described in supplementary methods; detailed primer sequences are shown in Table 1. GAPDH or 5S ribosomal RNA expression was used as internal control.

| Measurement of protein levels by Western blotting
Protein extracts from pancreas homogenates and acini normalized for protein concentration (10 µg protein) were subjected to SDS-PAGE and then transferred onto PVDF (Polyvinylidene fluoride) or nitrocellulose membranes (Thermo Fisher Scientific) for immunoblotting; a detailed protocol is described in supplementary methods.
Band intensities were quantified by densitometry using the Image J software (https://imagej.nih.gov/ij/) and normalized by loading control (amylase in case of acini and Ponceau staining for pancreas extracts).

| LC-MS/MS measurements and data analysis
A total of 7 μg of protein resuspended in 8 M urea/ 2 M thiourea was prepared and digested with trypsin using a bead-based SP3 protocol. 35 Details on mass spectrometric and data analysis are

| Statistical analyses
Data are presented as means ± SEM, and the number of independent experiments is indicated. Statistical analyses, significant when P < .05, were performed with GraphPad software (GraphPad) using one-way analyses of variance (ANOVA) followed by Turkey's multiple comparison test or unpaired Student's t test, as appropriate. For the Supplementary   Figure 1, SigmaPlot 11.0 (Systat Software) was used for the analyses.

| The β5i/LMP7 subunit is up-regulated in acute pancreatitis
First, we determined whether β5i/LMP7 subunit and its con- subunit most likely to compensate for decreased chymotrypsinlike activity ( Figure 1A). In a next step, we investigated whether the regulation of the β5i/LMP7 subunit would also be shifted in an in vivo mouse model of pancreatitis. We therefore used the caerulein model, which induces a mild and reversible form of the disease. 29 Three experimental groups were designed: control (0 hour), acute phase (8 hours) and recovery phase (24 hours).
Interestingly, the β5i/LMP7 transcript levels were significantly F I G U R E 1 Expression and regulation of LMP7 subunit in isolated acini and in experimental acute pancreatitis. A, Isolated acini were treated with 1 µM of CCK for 8 h and 24 h. The levels of LMP7 and β5 subunits were determined by Western blotting. Amylase was used as an internal control, and a ratio of LMP7 was calculated. B, Transcript levels of LMP7 were evaluated by real-time PCR in the pancreas after caerulein-induced pancreatitis. 5S was used as an internal control (n = 5-8). C, Western blotting analyses of LMP7 subunit in the pancreas during pancreatitis (n = 9-10). Data from independent experiments are expressed as means ± SEM. *P < .05, **P < .01, ***P < .001 its levels returned to baseline as measured in the control group ( Figure 1B). Importantly, no β5i/LMP7 transcripts were detected in the pancreas from β5i/LMP7 −/− mice. Western blot analysis of pancreas homogenates showed an approximately threefold increase in β5i/LMP7 protein levels at 24 hours compared to basal levels ( Figure 1C).

| Higher pancreatic damage in β5i/LMP7deficient mice
To investigate the role of the β5i/LMP7 immunoproteasome subunit in pancreatitis, suggested by its up-regulation in vivo, disease severity was evaluated in β5i/LMP7 −/− and littermate controls β5i/LMP7 +/+ mice. We performed the caerulein-induced pancreatitis model, including the same experimental groups as before.
Biochemical markers of pancreatic injury, 11 Figure 1). Thus, we showed that the lack of β5i/LMP7 did not alter the early protease activation within acini but during later pancreatitis progression.
Since intracellular protease activation is associated with acinar cell death, 30  Pancreatic injury was approximately twofold higher in the absence of β5i/LMP7 at the 24 hours time point ( Figure 2E). To assess cell death, we measured lactate dehydrogenase (LDH) activity in serum, which was also increased by 20% in β5i/LMP7 −/− mice at 8 hours ( Figure 2F). Of note, increased inflammatory cell infiltration was detected in the absence of β5i/LMP7 ( Figure 2E), leading to the next question to be addressed.

| β5i/LMP7 deficiency correlates with increased inflammation
Our group has previously shown that acinar cell damage is linked to a pro-inflammatory state characterized by increased infiltration of macrophages and neutrophils into the pancreas, which correlates with the extension of necrotic areas and decrease of healthy exocrine tissue. 34 We therefore next characterized these leukocytes population by immunofluorescence in pancreatic tissue from all experimental groups. This approach was performed using specific antibodies for profiling of pro-and anti-inflammatory macrophages phenotypes (CD68 and CD206, respectively) and neutrophils (Ly6g). No significant difference was noticed in the count of infiltrating macrophages ( Figure 3A). However, a significant increase in neutrophil population was observed at 24 hours ( Figure 3B). Next, we measured the myeloperoxidase (MPO) activity in lungs, which is a peroxidase enzyme abundantly expressed in neutrophil granulocytes. 36 As expected from the neutrophil infiltration in the absence of β5i/LMP7, we also detected enhanced MPO activity ( Figure 3B). We then quantified the activation of pro-inflammatory mediators in serum and in pancreas.
Along with extended acinar cell injury, we found that the transcription of pro-inflammatory cytokines IL-1β, CXCL-10 at 8 hours and IFNβ at 24 hours was increased in the absence of β5i/LMP7 compared to littermate controls by approximately 2.5-fold to sixfold ( Figure 3C). Likewise, serum levels of circulating IL-6 cytokine were elevated after 8 hours of pancreatitis ( Figure 3D). To analyse whether the alterations in the pro-inflammatory cytokines regulation originated from pancreatic acini or macrophage BMDM cells, acinar cells were either stimulated with CCK alone or co-incubated with BMDM cells. This second approach allows the evaluation of macrophages activation in the onset of pancreatitis. 34 In both cell types, we observed a higher regulation of IL-6 and IFNβ cytokines in the absence of β5i/LMP7. The CXCL-10 chemokine was up-regulated in pancreatic acini whereas IL-1β was significantly elevated in BMDM. As seen in isolated acini, β5i/LMP7 subunit was also induced in co-exposed acini macrophages (Supplementary  Figure 4D). These findings suggest that an impairment of immunoproteasome function has an impact on protein degradation during pancreatitis, which it is not due to proteasome .
In addition to the ubiquitin-dependent protein degradation pathway, we assessed whether the absence of β5i/LMP7 also had an influence on autophagy. Autophagy dysregulation has been identified as one key mechanism involved in the pathogenesis of pancreatitis. LC3-II is the lipidated form of its cytosolic counterpart (LC3-I) and is required for autophagosome formation. Sequentially, it engulfs cellular material for degradation and fuses with lysosomes, comprising the entity, in which protein cleavage occurs. 36 Analysis of LC3-II protein levels showed similar activation in the absence of β5i/LMP7, inferring that the induction of autophagy was not changed in our model (Supplementary

| Sustained modulation of the unfolded protein response in β5i/LMP7 −/− mice
The expected consequences of immunoproteasome impairment would be delayed protein degradation, leading to protein accumulation within the ER. 37 Thus, we hypothesized that immunoproteasome deficiency may result in increased ER stress and consequently, turn-on of the UPR signalling pathways. To address the UPR response, we performed Likewise to CCK treatment, we observed a significant twofold higher up-regulation of ER stress transcripts, such as BIP, ATF4 and CHOP, in the absence of the immunoproteasome (Supplementary Figure 4).
These observations confirm the importance of a functional immunoproteasome in regulating UPR induction.

| Decreased oxidative phosphorylation is associated with acinar cell death in β5i/LMP7 −/− mice
To characterize the protein content during acute pancreatitis, we performed mass spectrometry analysis using pancreatic insoluble fractions from littermate controls and β5i/LMP7 −/− mice.
Abundance ratio of samples from 8 hours vs 0 hour groups was calculated for each genotype and was plotted for selected proteins as a heat map ( Figure 6). We confirmed imbalanced protein homeostasis by decreased regulation of protein translation and folding in immunoproteasome-deficient mice. Moreover, we observed decreased abundance ratios of proteins from the oxidative phosphorylation metabolic pathway, suggesting ATP depletion and increased oxidative stress. A list of the proteins from the heat map can be found in Supplementary Table 1.
Overall, our data imply that the immunoproteasome plays a protective role in acinar cell homeostasis, since elevated cell death was observed in the absence of β5i/LMP7 in pancreatitis. The effect on pancreatic injury did not relate to changes in acinar cell regeneration, considering that the proliferation rate visualized by immunohistochemistry was not altered in the absence of β5i/LMP7 (Supplementary Figure 5). Therefore, the immunoproteasome appears to contribute to the clearance of ubiquitin-protein aggregates, resolution of inflammation and the reversal of UPR activation in pancreatitis.

| D ISCUSS I ON
In this study, we report a protective role of the immunoproteasome in pancreatitis and a beneficial effect on pancreatic injury resolution.
The expression of the immunoproteasome subunits is known to be regulated not only in immune cells but also in non-immune cells, for example in pancreatic beta cells, muscle cells and adipocytes. [38][39][40] Previous reports have demonstrated the importance of this F I G U R E 5 Stronger induction of ER stress transcripts in acute pancreatitis and isolated acini from LMP7 −/− mice. ER stress transcripts levels of BIP, ATF4 and sXBP-1 were determined by real-time qPCR using specific primers and 5S or GAPDH as housekeeping. A, RNA total from pancreas after pancreatitis; LMP7 +/+ (n = 10-13) and LMP7 −/− groups (n = 10-21). B, CHOP protein expression was assessed by Western blotting in the pancreas (n = 9-12). C, Isolated pancreatic acini stimulated with 1 µM of CCK for 6 h (n = 3). D, Co-culture of acini and macrophages cells for 6 h at 37℃ (n = 3-5). Data are illustrative of independent experiments and expressed as means ± SEM. *P < .05; **P < .01; ***P < .001 Notably, the β5i/LMP7 subunit but not the β5 constitutive subunit was up-regulated in an in vitro and an in vivo model of pancreatitis. Increased β5 protein levels in acini from β5i/LMP7-deficient mice were observed as a compensatory mechanism, as seen before in other cell types. 42,43 However, from the data obtained in this study, the immunoproteasome represents a major proteasome type involved in a rapid and efficient response to inflammation and ER stress during pancreatitis.
In the development of pancreatitis, well known as a primarily sterile disease, two major players have been described in the past. It is well established that acini, in response to supraphysiological secretagogue stimulation, develop ER stress and subsequently activate UPR. 51 In turn, UPR up-regulation has been recognized as a protective mechanism against acinar cell injury in pancreatitis. 14 The UPR restores ER protein homeostasis through inhibition of protein synthesis, enhancement of protein folding and degradation mechanisms, such as autophagy and the UPS. 17 Recently, the UPR and the immunoproteasome have appeared as pivotal networks in the maintenance of protein homeostasis and cell fate control. 52,53 In our work, we show that the immunoproteasome is crucial for limiting UPR activation in isolated pancreatic acini. The UPR encircles stress response signalling pathways triggered by accumulation of misfolded or damaged proteins in the ER lumen. 54 Although some studies did F I G U R E 6 Differential protein content from pancreatic insoluble fractions of littermate controls and LMP7 −/− mice. A total of 7 μg of protein resuspended in 8 M urea/ 2 M thiourea was prepared and digested with trypsin using a bead-based SP3 protocol. Measurements were performed on a QExactive HF mass spectrometer. Abundance ratios were calculated as the median of all possible pairwise peptide ratios calculated between replicates of all connected peptides. The heat map shows the significant abundance ratio from 8 h vs 0 h samples. Data represent independent experiments and are shown as means ± SEM. *P < .05. n = 3 (0 h), n = 4 (8 h) not report differences in the capacities of the constitutive-and the immunoproteasome to degrade ubiquitinated protein targets, 43,55 our work supports an essential role of the immunoproteasome in adjusting this process and synchronizing UPR activation in pancreatitis.
Interestingly, type I interferonopathies, characterized by excessive type I interferon (IFNα/β) production, have been associated with proteasome impairment. Mutations in PSMB8 encoding β5i/LMP7 cause a rare autoinflammatory disease (proteasome associated autoinflammatory syndromes, PRAAS), which is accompanied by transcription of IFN-stimulated genes, such as CXCL-10. In this scenario, β5i/LMP7 deficiency leads to an imbalance of the intracellular protein homeostasis, eliciting accumulation of ubiquitinated proteins and UPR activation, which is accompanied by a type I interferon signature. 56,57 Unbalanced UPR due to proteasome dysfunction has been implicated in the innate immune response, with the immunoproteasome playing a central role in the regulation of inflammation and UPR. The crosstalk between UPR and type I IFN signature has been proposed to evoke a sterile inflammation in the event of increased ER stress levels. [57][58][59][60] Our data align with these findings and provide further insight in the role of the immunoproteasome in acinar cell protein homeostasis and into the connection of inflammation and UPR in acute pancreatitis.
As a consequence of UPR dysregulation, represented by a prolonged time span for activation of sensor proteins and downstream targets, cell death appears to be the outcome. 53 Immunoproteasome dysfunction correlates with sustained pancreatic damage, which is accompanied by increased acinar cell death, namely apoptosis and necrosis. These observations support previous findings, implying that the UPR is to a certain extent protective via restoring protein homeostasis. Nevertheless, its sustained activation may lead to cell death. 18,61 One of the proposed mechanisms involves the ATF4-CHOP signalling pathway, a downstream target of the PERK stress sensor. In acute pancreatitis, C/EBP homologous protein (CHOP) is induced by ER stress and was linked to apoptosis by inflammasome activation and IL-1β secretion. 62 In addition, CHOP was also shown to regulate IL-6 transcription via STAT signalling, which leads to a sustained inflammation in case of protease impairment. 58 Consistent with these findings, the absence of the immunoproteasome links a stronger CHOP activation with higher apoptosis rates in the course of pancreatitis. The proteomic data show that impaired protein homeostasis and increased cell death are also correlated with diminished oxidative phosphorylation and protein translation/folding.
Thus, we identified a role of the immunoproteasome by responding efficiently to ER stress, reducing proteotoxic stress levels and restoring acinar cell viability.
Selective LMP7 inhibition has been suggested as a therapeutic modality in autoimmune diseases, such as rheumatoid arthritis, since it induces an anti-inflammatory response at lower doses than nonselective inhibitors, such as bortezomib, and reduces serum autoantibody levels, blocking disease progression. 47 In type I diabetes, abnormal self-antigen production leads to autoimmunity in patients and animal models. Due to its role in the generation of MHCpresented autoantigenic epitopes, the immunoproteasome might play a part in the autoimmune destruction of insulin-producing pancreatic β-cells. 38 In acute pancreatitis, the balance of UPR and the maintenance of acinar cell protein homeostasis are important cellular processes involved in the severity of tissue injury. The immunoproteasome appears to be a critical proteolytic complex in mitigating the pro-inflammatory response and preventing exacerbated pancreatic injury acute from pancreatitis. Future studies will be needed to identify pharmaceutically usable agents that improve or maintain the function of the immunoproteasome in order to translate this novel treatment target into a therapeutic modality.

ACK N OWLED G EM ENTS
This work was supported by the Center of Excellence Mecklenburg-Vorpommern (ESF/14-BM-A55-0045/16, PePPP). We would like to thank Davide Figini for assisting with imaging analyses, Christian Hentschker for the LC-MS/MS measurements, Franziska Thiel, Kathrin Gladrow and Jenny Radel for excellent technical assistance.

CO N FLI C T O F I NTE R E S T
The authors have declared that no conflict of interest exists.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.