PERK eIF2 alpha kinase is required to regulate the viability of the exocrine pancreas in mice
© Iida et al; licensee BioMed Central Ltd. 2007
Received: 06 April 2007
Accepted: 29 August 2007
Published: 29 August 2007
Deficiency of the PERK eIF2α kinase in humans and mice results in postnatal exocrine pancreatic atrophy as well as severe growth and metabolic anomalies in other organs and tissues. To determine if the exocrine pancreatic atrophy is due to a cell-autonomous defect, the Perk gene was specifically ablated in acinar cells of the exocrine pancreas in mice.
We show that expression of PERK in the acinar cells is required to maintain their viability but is not required for normal protein synthesis and secretion. Exocrine pancreatic atrophy in PERK-deficient mice was previously attributed to uncontrolled ER-stress followed by apoptotic cell death based on studies in cultured fibroblasts. However, we have found no evidence for perturbations in the endoplasmic reticulum or ER-stress and show that acinar cells succumb to a non-apoptotic form of cell death, oncosis, which is associated with a pronounced inflammatory response and induction of the pancreatitis stress response genes. We also show that mice carrying a knockout mutation of PERK's downstream target, ATF4, exhibit pancreatic deficiency caused by developmental defects and that mice ablated for ATF4's transcriptional target CHOP have a normal exocrine pancreas.
We conclude that PERK modulates secretory capacity of the exocrine pancreas by regulating cell viability of acinar cells.
Acinar cells of the exocrine pancreas synthesize and store copious quantities of the major digestive enzymes, poised to respond to episodic feeding events. Pro-enzymes (zymogens) are processed and folded in the rough endoplasmic reticulum and Golgi complex and are packaged into zymogen granules. Binding of a secretagogue such as cholecystokinin or acetylcholine to its receptors initiates Ca2+ signaling and results in regulated exocytosis of zymogens into the duct. However, abnormal signaling caused by various factors including hyper-stimulation with high doses of secretagogues can activate zymogens prematurely in the pancreas, which results in autodigestion of the cells and acute pancreatitis.
Previously we , and others  have independently showed that the lack of PERK (EIF2AK3) in mice results in atrophy of the exocrine pancreas. This is consistent with earlier findings of pancreatic insufficiency in Wolcott-Rallison syndrome, a rare autosomal recessive disorder caused by a mutation in the human Perk gene [3, 4]. PERK is an endoplasmic reticulum (ER)-resident kinase that phosphorylates eukaryotic translation initiation factor 2α (eIF2α) [5, 6]. Hyper-activation of PERK in cultured cells by pharmacological ER-stress inducers results in global repression of protein synthesis and translational induction of ATF4 [7–9]. This, in turn, induces the transcription of GADD153/CHOP. Consequently, the observed atrophy of the exocrine pancreas, as well as the loss of the insulin-secreting β-cells of the endocrine pancreas, in the PERK-deficient animals was interpreted as a defect in the unfolded protein response (UPR) which resulted in uncontrolled ER-stress, followed by apoptotic cell death [2, 10]. However, there is no direct evidence for uncontrolled ER-stress and apoptotic cell death in the exocrine pancreas of mice with knockout mutations of Perk (PKO), and recently the diabetes associated with PERK-deficiency in mice was shown to be caused by a proliferation and differentiation defect of the insulin secreting β-cells during fetal and neonatal development and not because of uncontrolled ER stress. A confounding problem in assessing the defects in the exocrine pancreas of PKO mice is that a multitude of severe neonatal defects occur in other organs and tissues prior to the detection of overt anomalies in the exocrine pancreas[1, 11]. These defects, which occur earlier in development, make it impossible to ascertain whether the anomalies occurring in the exocrine pancreas of PKO mice are cell-autonomous or secondary to dysfunctions in other organs.
To elucidate the mechanisms underlying the cell death and atrophy of the exocrine pancreas of the PKO mice, we have generated a pancreatic acinar cell-specific knockout mutation of the Perk gene (exPKO) in mice using the Cre-loxP system and have examined the critical physiological functions, the ER-stress pathway, and the mode of cell death in acinar cells.
PERK is cell-autonomously required in the exocrine pancreas
We generated pancreatic acinar cell-specific Perk knockout mice by crossing a mouse carrying Cre recombinase driven by the Elastase-1 promoter ("Ela-Cre")  to a mouse carrying a floxed Perk gene . The Elastase-1 promoter directs expression specifically in the rodent exocrine pancreas [12–14]. The tissue-specific knockout mice (exPKO; Ela-Cre h+; Perk flox/flox) showed a high CRE-mediated recombination rate (average 64.7% ± 0.06 [s.d.]) in the total pancreas at postnatal day 11 (P11). Since 7–8% of the total pancreas is comprised of other cell types we estimate that 70% of acinar cells contained the deletion of Perk. The fraction of phosphorylated eIF2α (to total eIF2α) in the exPKO pancreas is reduced to approximately 50% of wild-type at P42 but is somewhat higher than what is seen in PKO mice of the same age.
The progression and magnitude of the cellular and molecular abnormalities of the exocrine pancreas were very similar in the exPKO and PKO mice, even though 30% of acinar cells in exPKO were estimated to be wild-type at P11. Two possible explanations for the equally severe atrophy in the exPKO mice are that the remaining wild-type cells (at P11) eventually undergo Cre-mediated deletion or that the dying PERK-deficient cells induce cell death of the remaining wild-type cells.
Unlike the PKO mice that have severely reduced β-cell mass and become diabetic by P22, the islets of Langerhans of the exPKO mice remained morphologically normal (Figure 1F) and exhibited normal distribution of α-glucagon and β-insulin cells (not shown). In addition the exPKO mice never became diabetic and exhibited normal glucose clearance rates at P22, P162 and P231 when subjected to a glucose tolerance test indicating that islet function was not perturbed (not shown). This was surprising because few acinar cells remained in these animals to provide the structure and typical cellular context for the islets (Figure 1F). These data suggest that the lack of PERK in acinar cells specifically affects the exocrine pancreas and does not contribute to the diabetes that develops in the PKO mice. In addition, expression of a Perk transgene in β-cells in the PKO background (β-Perk; PKO; see ) rescues islet mass and diabetes but not the exocrine atrophy. Furthermore, tissue-specific Perk ablation in the entire pancreas (pcPKO, using Pdx1-Cre) causes both diabetes and exocrine atrophy; by contrast, mice with gene deletion only in islets (enPKO, using Ngn3-Cre) exhibit diabetes without exocrine atrophy . Thus PERK is required cell-autonomously in pancreatic acinar cells for their viability, and the multitude of other abnormalities of the PKO mice do not cause acinar cell death.
Acinar cell death in exPKO and PKO mice typically appeared around the neonatal-juvenile transition at postnatal 3 weeks. Numerous developmental and nutritional changes occur during this interval including a major transition in diet as pups are weaned from high-fat milk onto solid chow, which puts new demands on the exocrine pancreas. To test if the Perk mutant phenotype onset is affected by the dietary change we postponed weaning until after P21 when pups were past the age when the mutant phenotype is typically initiated. We found that delaying the dietary change that occurs at weaning does not prevent the progression of exocrine pancreatic atrophy in exPKO mice (not shown). To determine if the atrophy of the exocrine pancreas in PERK-deficient mice is developmentally regulated, we utilized a tamoxifen-inducible Cre-deletor strain targeted to the exocrine pancreas (Elastase-CreERT2) to conditionally mutate the Perk gene in adult mice. Multiple doses of tamoxifen were administered to 2–3 month-old mice bearing floxed alleles of Perk and the Elastase-CreERT2 transgene. Several acinar cells in the tamoxifen-treated mutant mice appeared degranulated, swollen and lacked normal nuclear content as observed in the PKO and exPKO mice (Figure 1H). As expected, the efficiency of CRE-mediated deletion of Perk was substantially lower than what was observed in the exPKO pancreas and consequently significantly fewer acinar cells succumbed to cell death. However, both gene deletion and the cell death were unique to mice carrying CreERT2; Perk flox/flox that were treated with tamoxifen and were absent in the tamoxifen-injected wild-type mice or non-injected controls. Therefore we conclude that PERK is required for maintaining the viability and function of the adult exocrine pancreas.
ExPKO acinar cells show pancreatitis-like phenotypes
Induction of the PSR genes was observed 1–2 days prior to detection of any other exocrine pancreatic anomalies, including other hallmarks of acute pancreatitis such as elevation of serum amylase, cell death, and leukocyte infiltration. During acute pancreatitis amylase is released into circulation as acinar cells break down and leak cellular contents into their surroundings. Thus the presence of amylase in serum is indicative of acute pancreatitis. Serum amylase was elevated significantly in exPKO mice as early as P18 (typically at P19; Figure 2B). Elevation of serum amylase was tightly correlated with an increase in the number of acinar cells undergoing cell death. As cell death progressed, however, serum amylase returned to low but detectable levels (by P40; data now shown). This is most likely due to catastrophic loss of acinar cells, which also occurs in severe cases of chronic pancreatitis [17, 18]. Serum amylase elevation and massive cell death were also accompanied by inflammatory responses. Gene profiling indicated the presence of an inflammatory response particularly at P19 and P32. Mutant samples showed a significant increase in genes predominately expressed in infiltrating macrophages and neutrophils such as Mac- 2 (a macrophage marker), Mmp12 (macrophage elastase), lysosomal proteins (Lip1 and Hexb), and Lcn2 (neutrophil gelatinase-associated lipocalin precursor) (see Additional file 1; also see Figure 2A for Mac-2).
To summarize, the exPKO exocrine pancreas exhibits characteristics of acute pancreatitis and inflammatory responses. The induction of the acute PSR genes occurs first in the exPKO mice, followed by the appearance of a series of pathologies including cell death, serum amylase elevation, and inflammation. These events immediately preceded an observed increase in the number of cells succumbing to cell death at the end of the third postnatal week.
ExPKO acinar cells show no evidence of ER-stress
Finally, in order to assess the level of ER-stress in acinar cells, the expression of the principal regulatory components and markers of ER-stress including BiP, Chop, and the spliced form of Xbp1 (Xbp1-s) were examined. The expression of these genes was not significantly altered in the total pancreas of exPKO mice. At P14-19, the average fold difference (± s.e.m., sample size) in mutants as compared to wild-type littermates was as follows: BiP, 1.21 (± 0.17, n = 7); Chop, 0.93 (± 0.22, n = 14); Xbp1-s, 1.15 (± 0.17, n = 11). This is consistent with microarray data that showed that expression of ER-resident and ER-stress-related genes was not significantly different between the mutant and wild-type pancreata at all ages examined (see Additional file 3), However, we did detect increases in mRNA levels of the pancreatic stress response (PSR) genes, which suggests that had ER-stress functions been similarly up-regulated we could have readily detected them as well.
ExPKO acinar cells die through oncosis, not apoptosis
The previously reported high frequency of TUNEL-positive acinar cells found in PKO mice as early as P18 led to the conclusion that the mode of cell death was apoptosis [1, 2]. However, it has become increasingly clear that the TUNEL assay cannot distinguish between apoptosis from other forms of cell death including oncosis [20–22]. The visual hallmarks of apoptosis are cellular shrinking and condensation of chromatin into apoptotic bodies, whereas oncotic cell death is characterized by cellular swelling, chromatin fragmentation without formation of apoptotic bodies, and loss of plasma membrane integrity. Necrosis has often been used synonymously with oncosis however it is now generally accepted that necrosis properly describes the morphological alterations that appear after cell death has occurred regardless of the mechanism . TEM analyses on the mutant pancreata failed to detect apoptotic bodies despite the appearance of numerous TUNEL-positive cells by P25 (see Figure 6B). Instead, acinar cells in the P25 mutant pancreas undergo vacuolation followed by karyolysis (Figure 5C–F) which are characteristic of oncosis (reviewed in ). Closer observation revealed two distinct TUNEL staining patterns. In both wild-type and mutant acinar cells a very small fraction (< 0.1%) of acinar cells exhibited intensely stained nuclei and fragmented nuclei, which are characteristic of apoptotic nuclear bodies. The fraction of cells exhibiting such TUNEL positive nuclei did not differ between genotypes. In contrast in exPKO acinar cells, a high frequency of TUNEL positive cells were observed to have diffuse staining throughout both the nucleus and cytoplasm (Figure 6A, B), consistent with the loss of nuclear membrane integrity. Furthermore, microarray analysis did not detect increased expression of apoptosis-related genes before, during, or after the peak period of cell death in the exocrine pancreas (P16, 19, and 32; see Additional file 4). During the peak period, up to 6% of the mutant acinar cells were TUNEL-positive and at this frequency apoptosis-related genes should have been readily detected had they been up-regulated in the exPKO pancreas. Instead, all array data indicated a significant induction of inflammatory genes in the mutant pancreas but not apoptotic genes (see Additional file 1). Of the genes that showed 2-fold or higher elevation in the P19 mutant pancreas, 38% were related to inflammatory, immune, or acute PSR responses.
An additional distinguishing characteristic of oncotic cells is the loss of plasma membrane integrity that ultimately leads to inflammatory responses [24, 25]. We developed a new technique to assess plasma membrane integrity in whole animal tissues based on the normal exclusion of the serum proteins apolipoprotein E, albumin, and IgG. In the normal exocrine pancreas, serum proteins are only found in the vasculature; however, in exPKO mice, serum proteins were detected in the degranulated acinar cells undergoing oncotic cell death (Figure 6C, D). Although serum proteins were abundantly detected in these cells, the mRNA levels were not elevated (not shown), thus eliminating the possibility of ectopic gene expression of these proteins. This observed loss of plasma membrane integrity as indicated by the passive uptake of serum proteins further substantiates our conclusion that the exPKO acinar cells succumb to oncosis. If apoptotic, the cells would have retained their plasma membrane integrity and been rapidly eliminated thus avoiding the induction of an acute inflammatory response. However, the degranulated acinar cells typically persisted for a few days before being phagocytosed, and the induction of an inflammatory response was clearly evident. Large numbers of infiltrating macrophages and neutrophils were detected in the exPKO pancreas by both acid phosphatase staining (Figure 6E, F), and by TEM (Figure 5D). Gene profiling also showed that transcripts for macrophage- and neutrophil-marker genes were elevated in the mutant pancreas (see Additional file 1). Phagocytosis of oncotic acinar cells by macrophages in mutant pancreata was also frequently observed by TEM (Figure 5D).
ATF4-deficient mice exhibit defects in the development of the exocrine pancreas
Gene expression in neonatal Atf4 KO mice as compared to wild-type littermates
CHOP upregulation is distal to both PERK and ATF4 activation and is generally considered to be pro-apoptotic, however, like caspase activation; it may also have a role in oncosis under certain circumstances . It was therefore of interest to see if the exocrine pancreas of Chop knockout mice bore any resemblance to mice deleted for either PERK or ATF4. The pancreata of adult Chop KO mice were found to be completely normal (not shown). To test for the presence of genetic interactions between Perk and its putative downstream targets, we generated Atf4+/-; Perk+/- and Chop+/-; Perk+/- double-heterozygotes. These mice exhibited normal exocrine and endocrine pancreata (not shown) thus failing to show a genetic interaction for pancreatic development and/or viability.
Although we show for the first time that the exocrine pancreas of the Atf4 KO is greatly reduced, the underlying cause appears to be unrelated to the exocrine pancreas defect in exPKO mice. The deficiency in exocrine pancreas mass in Atf4 KO mice was observed early in postnatal development as a consequence of hypotrophy whereas the exocrine insufficiency of the exPKO mice was caused by death of mature acinar cells resulting in atrophy and a robust inflammatory response.
PERK is required for the viability of pancreatic acinar cells
The generation of the exocrine pancreas-specific Perk-knockout mice enabled us to critically test whether exocrine pancreas defects seen in the global PKO mouse are the direct result of the loss of PERK expression in the acinar cells or are secondary to defects occurring earlier in other tissues or organs. We found that the exPKO mice had the same catastrophic loss of acinar cells as seen in global knockout mice, even though the deletion of Perk in the exPKO pancreas was incomplete (ca. 70%) at P11. We surmise that the nearly complete atrophy of the exocrine pancreas in exPKO is due to continued activity of the acinar-cell targeted Cre enzyme to delete the Perk gene in the remaining cells over the ensuing days or that the inflammatory response causes the death of the remaining wild-type cells as is seen when NFkappaB is overexpressed . The remarkable consistency in mutant pancreatic phenotype among all of the lines that are ablated for Perk in the exocrine pancreas is a strong indication that expression of PERK in acinar cells is required for their viability. Although acinar cell death and atrophy were the predominant outcome of ablating PERK in the exocrine pancreas, a subset of severely affected animals harbored some acini that had apparently dedifferentiated into duct-like structures. This phenomenon is also seen in experimentally-induced pancreatitis  and in animals that have a genetic block in TGFβ signaling [31, 32].
With the exception of the exocrine pancreatic deficiency, the exPKO mice did not display the other major defects associated with the Wolcott-Rallison syndrome in humans and the PKO mice. In some exPKO animals, the exocrine pancreas was replaced almost entirely by adipocytes, fibroblasts, and leukocytes, and the islets of Langerhans were no longer surrounded by acinar cells. These animals exhibited normal islet structure and glucose homeostasis. This demonstrates that the endocrine pancreas can survive and function without the exocrine pancreas being adjacent. It has been reported that chronic pancreatitis often results in diabetes, suggesting that dysfunction of the exocrine pancreas has deleterious effects on the endocrine pancreas (reviewed in ). In contrast, our findings suggest that the loss of the exocrine pancreas does not necessarily affect viability and function of the endocrine pancreas.
The onset of pancreatic atrophy in PERK-deficient mice occurs during the late neonatal stage of development and is first seen as sporadic degranulation of acinar cells followed by extensive cell death within a few days. Given the early postnatal onset of this atrophy, we considered whether these defects were dependent upon developmental age and nutrition. By delaying weaning, we showed that onset of exocrine pancreas atrophy in exPKO mice was not altered by changing the diet. In addition, we found that postponing the deletion of the Perk gene in the exocrine pancreas until the adult stage still lead to oncotic cell death, thus demonstrating that PERK is essential for the viability of the adult exocrine pancreas. In contrast, PERK is required for the fetal development and proliferation of the insulin-secreting β-cells of the endocrine pancreas, but is not required for β-cell viability in mature adults .
The mode of cell death in PERK-deficient exocrine pancreas is oncosis
The acinar cells of the exPKO and PKO mice exhibit a dramatic increase in cell death. Initially we  and others  had proposed that the form of cell death was apoptosis but this was based entirely upon the detection of DNA fragmentation in the nuclei of acinar cells using TUNEL immunohistochemistry. It has been reported that TUNEL does not distinguish between apoptosis and other forms of cell death such as oncosis. Therefore, our conclusion that exPKO acinar cells succumb to oncosis rather than apoptosis is based primarily on morphological criteria [20–22]. Oncotic cells are characterized by cell swelling, vacuolization, chromatin fragmentation without formation of apoptotic bodies, and increased membrane permeability [24, 25]. Fickert and colleagues  have recently shown that oncosis, not apoptosis, is the predominant form of cholestasis-induced cell death in the liver, and have provided a set of criteria for distinguishing between these forms of cell death in animal tissues that we have used herein.
Our conclusion that the mode of cell death in the PERK-deficient exocrine pancreas is oncosis derives from our key findings that include the appearance and progression of cell-swelling, karyolysis, and loss of plasma membrane integrity, as well as the induction of an inflammatory response and the absence of up-regulated apoptotic gene expression. Apoptotic cell death is often difficult to detect morphologically owing to the rapid cellular disintegration and immune clearance of the cells within a few hours. This rapid clearance and maintenance of plasma membrane integrity of apoptotic cells is thought to be a key factor limiting inflammation. In contrast, the oncotic acinar cells observed in the PERK-deficient exocrine pancreas seem to persist for a longer period as "ghost" cells that have lost plasma membrane integrity, and this may provide an opportunity for an inflammatory response to occur.
The exPKO exocrine pancreas exhibits normal protein synthesis and secretion without any indication of ER-stress
The major function of PERK has been proposed to be the repression of global protein synthesis during ER stress and the activation of translation of specific genes that alleviate ER stress. Genetic ablation of PERK was therefore predicted to cause uncontrolled ER stress in normal cells resulting in heightened ER dysfunction and cell death. The gross pancreatic abnormalities seen in PERK-deficient mice, including low insulin-secreting beta cell mass and exocrine pancreatic atrophy, were initially assumed to confirm this hypothesis [1, 2]. However the neonatal diabetes of PERK-deficient mice has been recently shown to be due insufficient proliferation and differentiation of the β-cells during fetal and neonatal development  and is not associated with either uncontrolled ER stress or cell death. In this study we also found that the exocrine pancreatic atrophy in PERK-deficient mice is inconsistent with the hypothesis that it is caused by uncontrolled or dysfunctional regulation of ER stress. First, protein synthesis and secretion are not significantly impaired in the mutant exocrine pancreas before or during the early phase of oncosis, although we assume that individual cells that have become degranulated and lost plasma membrane integrity are physiologically dysfunctional. Second, we did not observe an increase in expression of the major ER-stress genes. Instead, the genes induced in the mutant pancreas prior to overt cytological changes are the PSR genes. Third, knockout mutations in PERK's known downstream targets, Atf4 and Chop, do not result in cell death and atrophy of exocrine pancreas. Finally, the structure of the ER is normal in mutant pancreatic acinar cells prior to the onset of oncosis and does not exhibit distention as previously described . Huang and coworkers  recently reported that PERK-deficient pancreatic acinar cells exhibit normal carbachol-stimulated ER Ca2+ signaling as well as normal reloading of Ca2+ following the response. The ability of exogenously supplied IP3 to induce ER Ca2+ release was also normal. However, they did find that the periodicity of the Ca2+ response was altered in Perk-deficient acinar cells and demonstrated that mutants have greatly reduced ER- plasma membrane interactions. They attribute these differences between mutant and wild-type to the previously reported distention and fragmentation of the ER in PKO acinar cells , a finding that our data now refutes. Yet overall the results of this study indicated that the receptor-mediated exocrine pathways are functional in Perk-mutant acinar cells, which is consistent with our finding that secretagogue-stimulated protein synthesis and secretion are normal in PERK-deficient acinar cells.
While we did not find evidence that ER stress was associated with pancreatitis in PERK-deficient mice, other studies have shown that induction of acute pancreatitis in wild-type rodents will lead to a classic ER stress response [35, 36]. Caerulein hyperstimulation of acute pancreatitis in rats results in phosphorylation of eIF2α, and a reduction in total pancreatic protein synthesis , which is consistent with a potential role for PERK in mediating an ER stress response during an episode of acute pancreatitis. Another study using the arginine model of acute pancreatitis in rats found that all of the principle pathways of ER stress signaling including PERK activation and eIF2α phosphorylation were up-regulated and that characteristic ER stress-related apoptosis appeared shortly thereafter in the exocrine acini . The differences between these acute models and what we observe in the exPKO pancreas may reflect overall differences in how pancreatitis is initiated and progresses. The hallmarks of pancreatic dysfunction in Perk mutant mice including the appearance of acinar cell degranulation, up-regulation of PSR genes and elevation of serum amylase take at least 2 weeks to appear whereas changes in pancreatic morphology, gene expression and the systemic manifestations of acute pancreatitis occur within hours of induction. It therefore seems reasonable to postulate that the signaling mechanisms responsible for pancreatitis in PERK-deficient mice differ from those that regulate the acute form of the disease in which PERK may participate. Common to all of these models of pancreatitis is the induction of inflammation that occurs in response to acinar cell death. In the animal models of acute pancreatitis inflammation has been commonly attributed to acinar cell necrosis; however based upon cytological descriptions these cells may be more properly considered oncotic .
We propose that PERK is required for the expression of anti-oncotic or pro-survival factors that are specific to the exocrine pancreas. PERK resides as a transmembrane protein kinase with its activation domain in the ER lumen and catalytic domain in the cytoplasm. PERK's activity is dynamically regulated by association/disassociation of the ER chaperone protein BiP/GRP78, whose activity is regulated by Ca2+ as a cofactor and by ATP as a substrate. PERK is highly activated by pharmacological reagents that deplete the ER lumen of Ca2+, and we have shown that normal secretagogues known to mobilize ER Ca2+ can also activate PERK under physiological conditions (B. McGrath and D. R. Cavener, unpublished data). These data suggest that PERK may be a sensor for Ca2+ mobilization from the ER associated with secretory activity. We speculate that activation of PERK through normal secretory activity results in PERK-dependent regulation of anti-oncotic or pro-survival factors that maintain the viability of the acinar cells at adult stages. Necrosis is now generally regarded as the post-mortem state that results from uncontrolled cell death whether it arises from oncosis or as a secondary consequence of failure by the immune system to clear apoptotic cells in a timely fashion (secondary necrosis) . However, recent reports have suggested that necrosis may also participate in regulated cell death associated with tissue renewal, embryogenesis, and immune response . By analogy it is therefore, possible that a normal oncotic pathway exists in the exocrine pancreas and is misregulated in PERK-deficient mice.
PERK is cell-autonomously required for the viability of the exocrine acinar cells but is not required for their normal protein synthesis and secretory functions. Perk-deficient acinar cells are lost by an oncotic mechanism of cell death and not by apoptosis as was originally reported [1, 2]. Acinar death in Perk null mice is not associated with up-regulation of ER stress functions. Moreover mice carrying mutations in Atf4 or Chop which are known downstream targets of PERK signaling during ER stress do not recapitulate the Perk mutant exocrine phenotype.
All the procedures that involved animal subjects were approved by the Institutional Animal Care and Use Committee at the Pennsylvania State University to assure humane treatment of the animals. The knockout and floxed alleles of Perk in mice were previously described by us . The floxed Perk allele contains two loxP sites located in intronic regions flanking three internal exons encoding segments of the lumenal and transmembrane domains. Cre-mediated deletion of these exons creates a frameshift resulting in a null phenotype. Atf4 and Chop knockout mice were obtained from Dr. Tim Townes, University of Alabama, and Dr. David Ron, New York University, respectively. For developmental stage-specific deletion of the Perk gene, tamoxifen (SIGMA) solution (20 mg/ml in corn oil) was injected (i.p.) into the Elastase-CreERT2 mice (2–3 months old) at 100 μl/animal . To avoid tamoxifen-induced pancreatitis , the chemical was injected 12 times intermittently over one month. After 5 days from the last injection, the mice were sacrificed for further analyses.
For weaning-extension, two mating pairs per experiment (two females and one male) were set up in one large cage. The resulting litters were kept together with their dams until pups reached P10 when they were divided randomly into two cages: one with regular chow (as control) and the other without chow. Dams were switched between the two cages every 12 hours to let them access food. Water was freely accessible to mice in both cages. Conditions were maintained until the pups reached P25. Pups of the control group were weaned from the mother at P21, the normal age for weaning.
Paraffin-embedded tissue was prepared as described previously . Sections were made at 3–5 μ thickness (3 μ for serial sections). Fluorescent labeling was carried out using anti-ApoE (1:250; Santa Cruz) and an appropriate fluorescently-conjugated secondary antibody (1:500; Molecular Probes). Nuclei were labeled with DAPI (Molecular Probes). Cell proliferation was measured by in vivo labeling with 5-bromo-2'-deoxyuridine (BrdU; SIGMA). intraperitoneally injected at 50 mg/body weight (kg) Mice were sacrificed 24 hours post-injection and BrdU incorporation was detected using anti-BrdU (1:30; DAKO). DNA fragmentation was detected by TUNEL (T dT-mediated dU TP N ick-E nd L abeling; DeadEnd Fluorometric TUNEL system, Promega) following the manufacturer's protocol. A leukocyte marker, acid phosphatase, was detected using Naphthol AS-BI phosphoric acid solution (SIGMA). This assay was performed on frozen sections (10 μ), of tissues fixed in 4% phosphate-buffered paraformaldehyde for 2 hours at room temperature and then cryoprotected in 15% sucrose at 4°C overnight. Digital images were captured using a Nikon E1000 microscope and processed using the Image Pro (Phase 3 Imaging Systems) software. The background brightness and contrast were kept equivalent for all images that were directly compared so as not to distort any apparent differences. Samples for transmission electron microscopy (TEM) were prepared as described previously .
Serum samples were separated from cardiac blood and amylase activity was detected using Amylase Reagent (RAICHEM) according to the manufacturer's protocol. The fold difference was calculated by normalizing each mutant sample to the average of the wild-type samples within the same litter and the averages of the fold difference from multiple experiments is presented in Results.
Pancreata were excised from mice that had been fasted overnight, and placed into a Petri dish containing KRB-HEPES buffer, 5 mg/ml bovine serum albumin (BSA; SIGMA), and 100 μg/ml soybean trypsin inhibitor (STI; SIGMA). The KRB-HEPES buffer contained: Krebs-Ringer bicarbonate (KRB) buffer mix (SIGMA), 0.9 mM CaCl2, 20 mM HEPES, and essential amino acids (GIBCO). The KRB-HEPES buffer with BSA and STI was injected to the tissue, and lobules were excised under a dissecting microscope .
Protein synthesis and secretion experiments
Approximately 30 lobules per sample were deprived of methionine and cysteine (Met/Cys) for 10 minutes at 37°C in Met/Cys-free Dulbecco's Modified Eagle Medium, 20 mM HEPES, Aprotinin (25 μg/ml; SIGMA), STI (100 μg/ml), and BSA (5 mg/ml). The lobules were then labeled with [35S]Met/Cys (500 μCi/ml) at 37°C for 15 minutes. The reaction was stopped by the addition of concentrated non-radioactive Met/Cys solution (0.1 M each). After lobules were washed twice, chase buffer (KRB-HEPES buffer with BSA [5 mg/ml], STI [100 μg/ml], and Aprotinin [25 μg/ml]) and either carbachol (at 0.5 μM) or water (for control samples) was added. The samples were incubated in a shaking water bath (37°C; 120 rpm) and were oxygenated every 15–20 minutes. An aliquot of buffer was removed at different time points and later trichloroacetic acid (TCA)-precipitated using ProteoPrep Protein Precipitation Kit (SIGMA). At the end of the experiment, lobules were washed, homogenized, and total protein was TCA-precipitated. Precipitates were dissolved in 1% sodium dodecyl sulphate; 0.1 M Tris, pH 8.9, and the resident radioactivity were measured by scintillation counting.
The conditions used to measure incorporation of [35S]Met/Cys into total protein were the same as those used for the pulse-labeling experiment. All samples were TCA-precipitated, and the radioactivity was normalized to the protein concentration of the total lysate.
Measurements of secreted amylase were obtained from basal and carbachol-stimulated lobules subjected to non-radioactive "chase" conditions described above. In this case both buffer aliquots and lobule homogenate were assayed directly for amylase activity without prior TCA precipitation.
RNA and DNA preparation
RNA samples were prepared using RNeasy Mini kit (QIAGEN). To prevent RNA degradation, the pancreas was excised and immediately placed in a mortar containing liquid nitrogen. The frozen tissue was crushed into small pieces, and homogenized in RLT buffer using QIAshredder columns (QIAGEN). The rest of the procedure followed the manufacturer's protocol. DNA samples for Real-time PCR were prepared using DNeasy Tissue kit (QIAGEN), following the manufacturer's protocol.
Quantification of expression levels
Reverse transcription (RT) was carried out using M-MLV reverse transcriptase, buffer and Random Primers (Promega) according to the manufacturer's protocol. Real-time PCR of cDNA of all genes except Perk was carried out using qPCR Core Kit for SYBR Green I (Eurogentec). Reactions were amplified using the ABI Prism 7000 Sequence Detection System with the following program: 50°C (2 minutes), 95°C (10 minutes), 40 cycles of (95°C [15 seconds], 60°C [1 minute]). A linear amplification range was chosen for analyses using ABI Prism 7000 SDS Software (Applied Biosystems). The dissociation curves were monitored for the quality and specificity of amplification. Levels of Xbp1-s (spliced form) were normalized to Xbp1-t (total) levels. All other genes were normalized to the levels of tubulin. Fold-difference calculations were carried out within the same litter because variation among litters was too large to permit pooling across litters. The average of the fold-difference values across different litters from several replicate experiments is presented in Results. The primer sequences were as follows: Bip, 5'-ACCCTTACTCGGGCCAAATT-3', 5'-AGAGCGGAACAGGTCCATGT-3'; Chop, 5'-CCAACAGAGGTCACACGCAC-3', 5'-TGACTGGAATCTGGAGAGCGA-3'; Mac-2, 5'-GATGAAGAACCTCCGGGAAAT-3', 5'-GTTATGTCACCACTGATCCCCA-3'; Pap, 5'-CCAAAAGAGGCCTGGAGGAC-3', 5'-GCCTCAGCGCTATTGAGCAC-3'; Reg2, 5'-CTTCCCCTTGGCTGAAAAAGA-3', 5'-TCTGGGCAGTTGATTTTGGC-3'; Reg3g, 5'-ATGCTGCTCTCCTGCCTGAT-3', 5'-GGCAACTTCACCTTGCACCT-3'; tubulin alpha, 5'-CTTGGAACCCACGGTCATC-3', 5'-AAGAGCTGGCGGTAGGTGC-3'; Xbp1-s, 5'-GAGTCCGCAGCAGGTG-3', 5'-GTGTCAGAGTCCATGGGA-3'; Xbp1-t, 5'-CACCTTCTTGCCTGCTGGAC-3', 5'-GGGAGCCCTCATATCCACAGT-3'. Primer sequences for Xbp1-s were previously published .
Real-time PCR for the Perk gene was carried out as described previously . The Perk "deleted region" primers amplify only the wild-type allele, whereas the Perk "total" primers amplify both wild-type and mutant alleles (Perk "deleted region", 5'-CAAACAAACCCAGCACCTTT-3', 5'-CATGGCCTGGCTTATTCTGT-3'; Perk "total", 5'-CATTCCTGGAACTGGGAGAA-3', 5'-CTCCATGGGACTCCTGGTTA-3'). The reactions were amplified using the GeneAmp 5700 DNA amplification system and analyzed with the GeneAmp 5700 SDS software (Applied Biosystems). To calculate the percentage of wild-type Perk allele remaining in the exPKO mice, the wild-type copy number estimated for the amplified "deleted region" was normalized to the copy number estimated using the "total" primers. Genomic DNA from wild-type, heterozygous and PKO were amplified simultaneously and used as standards to estimate the deletion frequency within the total exPKO pancreas. Theoretically, the percentage of wild-type allele left for the three standards should be approximately 100%, 50% and 0% respectively and this is what was observed.
All procedures followed the Affymetrix protocol. GeneChip Mouse Expression Array 430A was used for all samples. Affymetrix Microarray Suite 5.1 was used to calculate a difference in gene expression between exPKO and wild-type at P19 and P32. Detailed methods for sample preparation and determination of expression differences are included in Additional Data. Microarray results were deposited to the Gene Expression Omnibus database Series #: GSE4422, 4423, 4424; Sample #: GSM99695, 99696, 99698, 99716, 99762, 99776, 99777, 99780.
We thank Dr. Robert C. De Lisle (University of Kansas School of Medicine) for his advise on lobule isolation and secretion experiments, Dr. D. A. Stoffers (University of Pennsylvania) for providing the Elastase-CreERT2 strain, Dr. Eric Sandgren (University of Wisconsin) for providing the Elastase-Cre strain, Dr. Tim Townes (University of Alabama) for providing the Atf4 knockout strain, and Dr. David Ron (New York University) for providing the Chop knockout strain.
This work was support by NIH grants DK62049 and GM56957 and the Pennsylvania Department of Health using Tobacco Settlement Funds.
- Zhang P, McGrath B, Li S, Frank A, Zambito F, Reinert J, Gannon M, Ma K, McNaughton K, Cavener DR: The PERK eukaryotic initiation factor 2 alpha kinase is required for the development of the skeletal system, postnatal growth, and the function and viability of the pancreas. Mol Cell Biol. 2002, 22 (11): 3864-3874. 10.1128/MCB.22.11.3864-3874.2002.PubMed CentralView ArticlePubMed
- Harding HP, Zeng H, Zhang Y, Jungries R, Chung P, Plesken H, Sabatini DD, Ron D: Diabetes mellitus and exocrine pancreatic dysfunction in perk-/- mice reveals a role for translational control in secretory cell survival. Mol Cell. 2001, 7 (6): 1153-1163. 10.1016/S1097-2765(01)00264-7.View ArticlePubMed
- Delepine M, Nicolino M, Barrett T, Golamaully M, Lathrop GM, Julier C: EIF2AK3, encoding translation initiation factor 2-alpha kinase 3, is mutated in patients with Wolcott-Rallison syndrome. Nat Genet. 2000, 25 (4): 406-409. 10.1038/78085.View ArticlePubMed
- Castelnau P, Le Merrer M, Diatloff-Zito C, Marquis E, Tete MJ, Robert JJ: Wolcott-Rallison syndrome: a case with endocrine and exocrine pancreatic deficiency and pancreatic hypotrophy. Eur J Pediatr. 2000, 159 (8): 631-633. 10.1007/PL00008394.View ArticlePubMed
- Shi Y, Vattem KM, Sood R, An J, Liang J, Stramm L, Wek RC: Identification and characterization of pancreatic eukaryotic initiation factor 2 alpha-subunit kinase, PEK, involved in translational control. Mol Cell Biol. 1998, 18 (12): 7499-7509.PubMed CentralView ArticlePubMed
- Harding HP, Zhang Y, Ron D: Protein translation and folding are coupled by an endoplasmic-reticulum-resident kinase. Nature. 1999, 397 (6716): 271-274. 10.1038/16729.View ArticlePubMed
- Harding HP, Novoa I, Zhang Y, Zeng H, Wek R, Schapira M, Ron D: Regulated translation initiation controls stress-induced gene expression in mammalian cells. Mol Cell. 2000, 6 (5): 1099-1108. 10.1016/S1097-2765(00)00108-8.View ArticlePubMed
- Lu PD, Harding HP, Ron D: Translation reinitiation at alternative open reading frames regulates gene expression in an integrated stress response. J Cell Biol. 2004, 167 (1): 27-33. 10.1083/jcb.200408003.PubMed CentralView ArticlePubMed
- Harding HP, Zhang Y, Bertolotti A, Zeng H, Ron D: Perk is essential for translational regulation and cell survival during the unfolded protein response. Mol Cell. 2000, 5 (5): 897-904. 10.1016/S1097-2765(00)80330-5.View ArticlePubMed
- Harding HP, Ron D: Endoplasmic reticulum stress and the development of diabetes: a review. Diabetes. 2002, 51 Suppl 3: S455-61. 10.2337/diabetes.51.2007.S455.View ArticlePubMed
- Zhang W, Feng D, Li Y, Iida K, McGrath B, Cavener DR: PERK EIF2AK3 control of pancreatic beta cell differentiation and proliferation is required for postnatal glucose homeostasis. Cell Metab. 2006, 4 (6): 491-497. 10.1016/j.cmet.2006.11.002.View ArticlePubMed
- Grippo PJ, Nowlin PS, Cassaday RD, Sandgren EP: Cell-specific transgene expression from a widely transcribed promoter using Cre/lox in mice. Genesis. 2002, 32 (4): 277-286. 10.1002/gene.10080.View ArticlePubMed
- Swift GH, Hammer RE, MacDonald RJ, Brinster RL: Tissue-specific expression of the rat pancreatic elastase I gene in transgenic mice. Cell. 1984, 38 (3): 639-646. 10.1016/0092-8674(84)90258-7.View ArticlePubMed
- Ornitz DM, Palmiter RD, Hammer RE, Brinster RL, Swift GH, MacDonald RJ: Specific expression of an elastase-human growth hormone fusion gene in pancreatic acinar cells of transgenic mice. Nature. 1985, 313 (6003): 600-602. 10.1038/313600a0.View ArticlePubMed
- Li Y, Iida K, O'Neil J, Zhang P, Li S, Frank A, Gabai A, Zambito F, Liang SH, Rosen CJ, Cavener DR: PERK eIF2alpha kinase regulates neonatal growth by controlling the expression of circulating insulin-like growth factor-I derived from the liver. Endocrinology. 2003, 144 (8): 3505-3513. 10.1210/en.2003-0236.View ArticlePubMed
- Iovanna J, Orelle B, Keim V, Dagorn JC: Messenger RNA sequence and expression of rat pancreatitis-associated protein, a lectin-related protein overexpressed during acute experimental pancreatitis. J Biol Chem. 1991, 266 (36): 24664-24669.PubMed
- Clavien PA, Robert J, Meyer P, Borst F, Hauser H, Herrmann F, Dunand V, Rohner A: Acute pancreatitis and normoamylasemia. Not an uncommon combination. Ann Surg. 1989, 210 (5): 614-620. 10.1097/00000658-198911000-00008.PubMed CentralView ArticlePubMed
- Yadav D, Agarwal N, Pitchumoni CS: A critical evaluation of laboratory tests in acute pancreatitis. Am J Gastroenterol. 2002, 97 (6): 1309-1318. 10.1111/j.1572-0241.2002.05766.x.View ArticlePubMed
- Huang G, Yao J, Zeng W, Mizuno Y, Kamm KE, Stull JT, Harding HP, Ron D, Muallem S: ER stress disrupts Ca2+-signaling complexes and Ca2+ regulation in secretory and muscle cells from PERK-knockout mice. J Cell Sci. 2006, 119 (Pt 1): 153-161. 10.1242/jcs.02731.View ArticlePubMed
- Dong Z, Saikumar P, Weinberg JM, Venkatachalam MA: Internucleosomal DNA cleavage triggered by plasma membrane damage during necrotic cell death. Involvement of serine but not cysteine proteases. Am J Pathol. 1997, 151 (5): 1205-1213.PubMed CentralPubMed
- Grasl-Kraupp B, Ruttkay-Nedecky B, Koudelka H, Bukowska K, Bursch W, Schulte-Hermann R: In situ detection of fragmented DNA (TUNEL assay) fails to discriminate among apoptosis, necrosis, and autolytic cell death: a cautionary note. Hepatology. 1995, 21 (5): 1465-1468. 10.1002/hep.1840210534.PubMed
- Fickert P, Trauner M, Fuchsbichler A, Zollner G, Wagner M, Marschall HU, Zatloukal K, Denk H: Oncosis represents the main type of cell death in mouse models of cholestasis. J Hepatol. 2005, 42 (3): 378-385. 10.1016/j.jhep.2004.10.016.View ArticlePubMed
- Van Cruchten S, Van Den Broeck W: Morphological and biochemical aspects of apoptosis, oncosis and necrosis. Anat Histol Embryol. 2002, 31 (4): 214-223. 10.1046/j.1439-0264.2002.00398.x.View ArticlePubMed
- Majno G, Joris I: Apoptosis, oncosis, and necrosis. An overview of cell death. Am J Pathol. 1995, 146 (1): 3-15.PubMed CentralPubMed
- Proskuryakov SY, Konoplyannikov AG, Gabai VL: Necrosis: a specific form of programmed cell death?. Exp Cell Res. 2003, 283 (1): 1-16. 10.1016/S0014-4827(02)00027-7.View ArticlePubMed
- Hettmann T, Barton K, Leiden JM: Microphthalmia due to p53-mediated apoptosis of anterior lens epithelial cells in mice lacking the CREB-2 transcription factor. Dev Biol. 2000, 222 (1): 110-123. 10.1006/dbio.2000.9699.View ArticlePubMed
- Masuoka HC, Townes TM: Targeted disruption of the activating transcription factor 4 gene results in severe fetal anemia in mice. Blood. 2002, 99 (3): 736-745. 10.1182/blood.V99.3.736.View ArticlePubMed
- Copple BL, Rondelli CM, Maddox JF, Hoglen NC, Ganey PE, Roth RA: Modes of cell death in rat liver after monocrotaline exposure. Toxicol Sci. 2004, 77 (1): 172-182. 10.1093/toxsci/kfh011.View ArticlePubMed
- Chen X, Ji B, Han B, Ernst SA, Simeone D, Logsdon CD: NF-kappaB activation in pancreas induces pancreatic and systemic inflammatory response. Gastroenterology. 2002, 122 (2): 448-457. 10.1053/gast.2002.31060.View ArticlePubMed
- Reid LE, Walker NI: Acinar cell apoptosis and the origin of tubular complexes in caerulein-induced pancreatitis. Int J Exp Pathol. 1999, 80 (4): 205-215. 10.1046/j.1365-2613.1999.00116.x.PubMed CentralView ArticlePubMed
- Logsdon CD, Keyes L, Beauchamp RD: Transforming growth factor-beta (TGF-beta 1) inhibits pancreatic acinar cell growth. Am J Physiol. 1992, 262 (2 Pt 1): G364-8.PubMed
- Bottinger EP, Jakubczak JL, Roberts IS, Mumy M, Hemmati P, Bagnall K, Merlino G, Wakefield LM: Expression of a dominant-negative mutant TGF-beta type II receptor in transgenic mice reveals essential roles for TGF-beta in regulation of growth and differentiation in the exocrine pancreas. Embo J. 1997, 16 (10): 2621-2633. 10.1093/emboj/16.10.2621.PubMed CentralView ArticlePubMed
- Gorelick FS: Diabetes mellitus and the exocrine pancreas. Yale J Biol Med. 1983, 56 (4): 271-275.PubMed CentralPubMed
- Zhang W: PERK eIF2AK3 control of pancreatic beta cell differentiation and proliferation is required for postnatal glucose homeostasis. Cell Metabolism. 2006, 4: 1-7. 10.1016/j.cmet.2006.11.002.View Article
- Kubisch CH, Sans MD, Arumugam T, Ernst SA, Williams JA, Logsdon CD: Early activation of endoplasmic reticulum stress is associated with arginine-induced acute pancreatitis. Am J Physiol Gastrointest Liver Physiol. 2006, 291 (2): G238-45. 10.1152/ajpgi.00471.2005.View ArticlePubMed
- Sans MD, DiMagno MJ, D'Alecy LG, Williams JA: Caerulein-induced acute pancreatitis inhibits protein synthesis through effects on eIF2B and eIF4F. Am J Physiol Gastrointest Liver Physiol. 2003, 285 (3): G517-28.View ArticlePubMed
- Bhatia M: Apoptosis versus necrosis in acute pancreatitis. Am J Physiol Gastrointest Liver Physiol. 2004, 286 (2): G189-96. 10.1152/ajpgi.00304.2003.View ArticlePubMed
- Restifo NP: Building better vaccines: how apoptotic cell death can induce inflammation and activate innate and adaptive immunity. Curr Opin Immunol. 2000, 12 (5): 597-603. 10.1016/S0952-7915(00)00148-5.PubMed CentralView ArticlePubMed
- Metzger D, Chambon P: Site- and time-specific gene targeting in the mouse. Methods. 2001, 24 (1): 71-80. 10.1006/meth.2001.1159.View ArticlePubMed
- Elisaf MS, Nakou K, Liamis G, Pavlidis NA: Tamoxifen-induced severe hypertriglyceridemia and pancreatitis. Ann Oncol. 2000, 11 (8): 1067-1069. 10.1023/A:1008309613082.View ArticlePubMed
- Scheele GA, Palade GE: Studies on the guinea pig pancreas. Parallel discharge of exocrine enzyme activities. J Biol Chem. 1975, 250 (7): 2660-2670.PubMed
- Back SH, Schroder M, Lee K, Zhang K, Kaufman RJ: ER stress signaling by regulated splicing: IRE1/HAC1/XBP1. Methods. 2005, 35 (4): 395-416. 10.1016/j.ymeth.2005.03.001.View ArticlePubMed
- Gene Expression Omnibus Database . [http://www.ncbi.nlm.nih.gov/geo/]
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.