Cluster of Differentiation 38 (CD38) Mediates Bile Acid-induced Acinar Cell Injury and Pancreatitis through Cyclic ADP-ribose and Intracellular Calcium Release*

Background: Bile acids cause ryanodine receptor (RyR) Ca2+ release and lead to injury in pancreatic acinar cells, yet the mechanism is unknown. Results: Inhibition of the RyR activator cADPR reduces bile acid-induced acinar cell Ca2+ release, cell injury, and pancreatitis. Conclusion: CD38-cADPR facilitates bile-induced Ca2+ release, cell injury, and pancreatitis. Significance: The CD38-cADPR pathway may serve as a target for the treatment of some forms of pancreatitis. Aberrant Ca2+ signals within pancreatic acinar cells are an early and critical feature in acute pancreatitis, yet it is unclear how these signals are generated. An important mediator of the aberrant Ca2+ signals due to bile acid exposure is the intracellular Ca2+ channel ryanodine receptor. One putative activator of the ryanodine receptor is the nucleotide second messenger cyclic ADP-ribose (cADPR), which is generated by an ectoenzyme ADP-ribosyl cyclase, CD38. In this study, we examined the role of CD38 and cADPR in acinar cell Ca2+ signals and acinar injury due to bile acids using pharmacologic inhibitors of CD38 and cADPR as well as mice deficient in Cd38 (Cd38−/−). Cytosolic Ca2+ signals were imaged using live time-lapse confocal microscopy in freshly isolated mouse acinar cells during perifusion with the bile acid taurolithocholic acid 3-sulfate (TLCS; 500 μm). To focus on intracellular Ca2+ release and to specifically exclude Ca2+ influx, cells were perifused in Ca2+-free medium. Cell injury was assessed by lactate dehydrogenase leakage and propidium iodide uptake. Pretreatment with either nicotinamide (20 mm) or the cADPR antagonist 8-Br-cADPR (30 μm) abrogated TLCS-induced Ca2+ signals and cell injury. TLCS-induced Ca2+ release and cell injury were reduced by 30 and 95%, respectively, in Cd38-deficient acinar cells compared with wild-type cells (p < 0.05). Cd38-deficient mice were protected against a model of bile acid infusion pancreatitis. In summary, these data indicate that CD38-cADPR mediates bile acid-induced pancreatitis and acinar cell injury through aberrant intracellular Ca2+ signaling.

In this study, we investigated the role of bile acids in mediating intracellular Ca 2ϩ release through a CD38-cADPR-dependent pathway. Bile acid exposure is known to cause acinar cell injury, and it may account for biliary pancreatitis, the most common etiology of acute pancreatitis in both children and adults (27). Bile acids exert their injurious effects on the acinar cell through Ca 2ϩ -dependent pathways. Specifically, bile acids trigger Ca 2ϩ release from endoplasmic reticulum and vesicular Ca 2ϩ stores. This occurs through activation of RyRs and inositol 1,4,5-trisphosphate receptors (5,18,28) and subsequent opening of store-operated Ca 2ϩ entry channels (7,16,29).
In a recent study by Cosker et al. (30), CD38 was shown to play a role in pancreatic acinar cell Ca 2ϩ signaling. Although another enzymatic product of CD38, nicotinic acid adenine dinucleotide phosphate (NAADP), was the focus of this study, cADPR levels were also reduced at base line in Cd38 Ϫ/Ϫ acinar cells compared with wild-type cells, and importantly, cholecystokinin stimulation did not induce an increase in cADPR levels, as it did in the wild-type cells. To exclude the effects of Ca 2ϩ influx, Cd38 Ϫ/Ϫ acini were stimulated in the absence of extracellular Ca 2ϩ and failed to evoke a Ca 2ϩ transient. On the basis of this work and the previous studies implicating RyR Ca 2ϩ signaling in bile acid-mediated pathology, in this study, we examined the role of CD38 and cADPR in intracellular Ca 2ϩ release, acinar cell injury, and pancreatitis due to bile acid exposure.

EXPERIMENTAL PROCEDURES
Reagents and Animals-All reagents were purchased from Sigma-Aldrich unless stated otherwise. Male Swiss Webster mice (weighing 20 -25 g; Charles River Laboratories, Wilmington, MA) were fed standard laboratory chow, given free access to water, and randomly assigned to control or experimental groups. Cd38 Ϫ/Ϫ mice were generated by one of us (F. E. L.) (31). Age-, sex-, and strain-matched control mice (C57BL/6 mice, The Jackson Laboratory) were used as wild-type controls. All animal experiments were performed using a protocol approved by the University of Pittsburgh Institutional Animal Care and Use Committee.
Preparation of Pancreatic Acini for Ca 2ϩ Imaging-Groups of pancreatic acinar cells were isolated as described previously (10) with minor modifications. Briefly, the pancreas was removed and then minced for 5 min in buffer containing 20 mM HEPES (pH 7.4), 95 mM NaCl, 4.7 mM KCl, 0.6 mM MgCl 2 , 1.3 mM CaCl 2 , 10 mM glucose, and 2 mM glutamine plus 1% BSA, 1ϫ Gibco minimum Eagle's medium nonessential amino acids (Invitrogen), 200 units/ml type 4 collagenase (Worthington), and 1 mg/ml soybean trypsin inhibitor. The tissue was incubated for 30 min at 37°C with shaking at 90 rpm. The digest was transferred to a 15-ml conical tube and washed with collagenase-free buffer. The suspension was vigorously shaken for 15-20 s to separate the cells into smaller clusters.
Detection and Analysis of Cellular Ca 2ϩ Signals-Acinar cells were loaded at room temperature with the high-affinity Ca 2ϩ -sensing dye Fluo-4/AM (K d ϭ 300 nM; Invitrogen). Acinar cells were plated on acid-washed glass coverslips and then mounted on a perifusion chamber. They were stimulated at room temperature with the bile acid taurolithocholic acid 3-sulfate (TLCS) at the concentrations indicated. A Zeiss LSM 710 laser scanning confocal microscope was used with a 40ϫ 1.4-numerical aperture objective. The dye was excited at a wavelength of 488 nm, and emission signals of Ͼ515 nm were collected every 2.5 s. Fluorescence from individual acinar cells was recorded. Recordings were analyzed using NIH ImageJ software, and mean fluorescence over time in each region was graphed.
Preparation of Pancreatic Acini-Groups of pancreatic acinar cells were isolated as described previously (32) with minor modifications. Briefly, the pancreas was removed and minced for 5 min in 1ϫ DMEM/nutrient mixture F-12 without phenol red (Invitrogen) plus 0.1% BSA and 2 mg/ml type-4 collagenase. The suspension was briefly incubated for 5 min at 37°C with shaking at 90 rpm. The buffer was removed, replaced with new collagenase buffer, and then incubated for 35 min. The suspension was filtered through a 300-m mesh (Sefar American, Depew, NY) and washed three times with collagenase-free buffer. Acinar cells were allowed to equilibrate for 5 min at 37°C prior to use.
Cell Injury Assays-Acinar cell injury was measured using a cytotoxicity assay for lactate dehydrogenase (LDH) leakage (Promega). Absorbance was measured at 492 nm. Results are expressed as percent LDH released into the medium. For propidium iodide uptake, acinar cells were incubated in a 48-well plate with 50 g/ml propidium iodide (Sigma) for 30 min prior to addition of 500 M TLCS. Fluorescence was measured at 536-nm excitation and 617-nm emission wavelengths. Data were normalized to total DNA by again measuring fluorescence after cell lysis with 0.5% Triton X-100.
Preparation of Human Acinar Cells-Pancreatic tissue was harvested from cadaveric donors as described by Bottino et al. (33). Briefly, specimens were transported in cold preservation fluid (histidine/tryptophan/ketoglutarate) with a cold ischemia time of 13 h. Fat, connective tissue, and blood vessels were trimmed away. The pancreas was washed with a mixture of antibiotics and then cut at the level of the neck to reveal the pancreatic duct. Catheters were placed in both sides of the transected duct, and a blend of exogenous enzymes, including collagenases and neutral proteases (GMP-grade, Serva, Heidelberg, Germany), freshly dissolved in Hanks' balanced salt solution was prewarmed to 28 -30°C and introduced intraductally. The pancreatic organ was then transferred to a Ricordi digestion chamber, and the pancreatic tissue was mechanically disrupted as described by Ricordi (34). Pancreatic cells were washed several times with cold RPMI 1640 medium supplemented with human serum albumin (2.5% total volume). Endocrine cell contamination was Ͻ1%. Acinar cells were kept in calcium-and magnesium-free Hanks' buffer, and cell injury assays were performed as described above.
Enzyme Activity Assays-Protease activity assays were performed at room temperature using fluorogenic substrates as described previously (35) with modifications. Briefly, 50 l of 400 M enzyme substrate were added to each homogenized sample, and the accumulation of fluorescence was measured over 12 min using a Synergy H1 fluorescence plate reader (BioTek, Winooski, VT) at 380-nm excitation and 440-nm emission wavelengths. The trypsin substrate was supplied by Peptides International (Louisville, KY) and had the amino acid sequence t-butoxycarbonyl-Gln-Ala-Arg-7-amino-4-methylcoumarin. The chymotrypsin substrate was supplied by Calbiochem and had the amino acid sequence succinyl-Ala-Ala-Pro-Phe-7-amino-4-methylcoumarin. Zymogen activity was normalized to total protein content.
Intraductal Bile Acid Infusion Model of Pancreatitis-Pancreatitis was induced by retrograde infusion of the bile acid TLCS (3 mM) into the distal common bile duct and pancreatic duct as described recently (36). Briefly, C57BL/6 mice between 8 and 12 weeks of age were anesthetized with isoflurane. A ventral incision was made to reveal the abdominal cavity. The duodenum was flipped to reveal its distal side and held in place by ligatures. The bile duct was identified, and a 30-gauge needle was inserted through the antimesenteric aspect of the duodenum to cannulate the biliopancreatic duct. TLCS was infused at 10 l/min for 5 min using a P33 perfusion pump (Harvard Apparatus, Holliston, MA). The exterior wound was closed using 7-mm wound clips, and a single injection of buprenorphine (0.075 mg/kg) was given immediately after the surgery. Normal saline-infused animals served as sham controls. Animals were allowed to recover on a heating pad for 90 min after the procedure. Mice were euthanized 24 h after induction.
Tissue Preparation and Histological Grading-The pancreas, duodenum, and spleen were fixed at room temperature for 24 h in 4% paraformaldehyde solution and transferred to 70% ethanol. Paraffin-embedded sections were stained with hematoxylin and eosin and graded using a 40ϫ objective over 10 separate fields in a blinded fashion. Pancreatic tissue was graded for edema, acinar cell vacuole formation, inflammation, and necrosis as described by Wildi et al. (37).   SEPTEMBER 20, 2013 • VOLUME 288 • NUMBER 38

RESULTS
Pharmacologic Inhibition of CD38 Attenuates Bile Acid-induced Ca 2ϩ Signals-We used TLCS to examine the effects of bile acids on acinar cell Ca 2ϩ release for two primary reasons. First, TLCS induces Ca 2ϩ signals at submillimolar concentrations below the critical micellar concentration. Second, it is the least hydrophilic and thus most potent of the naturally occurring bile acids (38).
To examine intracellular Ca 2ϩ release and to exclude the influence of extracellular Ca 2ϩ , acini were loaded with the highaffinity Ca 2ϩ dye Fluo-4/AM (Fig. 1A) and perifused in a nominally Ca 2ϩ -free medium. Using time-lapse laser scanning confocal microscopy, we observed that perifusion with 500 M TLCS caused Ca 2ϩ oscillations in over 90% of the acini (Fig. 1B). Lower concentrations of TLCS (50 M) in the absence of Ca 2ϩ did not induce a Ca 2ϩ transient (data not shown).
To determine whether TLCS-induced Ca 2ϩ release is dependent on CD38, acinar cells were first perifused with TLCS (500 M) and then co-treated with nicotinamide (20 mM). We found that nicotinamide abolished the Ca 2ϩ response to TLCS (Fig. 1, C and D). To determine whether the effect was reversible, TLCS was administered after washing off nicotinamide, and a second peak was observed. Similar results were observed when acinar cells were pretreated with nicotinamide prior to TLCS stimulation (Fig. 1, E and F). The Ca 2ϩ transient observed following withdrawal of nicotinamide was a global surge, which appears as a single Ca 2ϩ spike.
Genetic Deletion of Cd38 Attenuates Bile Acid-induced Ca 2ϩ Signals-In addition to inhibiting CD38, nicotinamide has several nonspecific effects. It can inhibit poly(ADP-ribose) polymerase (39) and sirtuins (40) and can scavenge reactive oxygen species (41). Therefore, to complement the pharmacologic data, we isolated acinar cells from Cd38-deficient mice (Cd38 Ϫ/Ϫ ) and stimulated them with TLCS (500 M). These mice have no pancreatic defect at base line and no gross phenotypic differences compared with wild-type mice (30,31). We observed that compared with wild-type cells, Cd38 Ϫ/Ϫ acinar cells exhibited a 30% reduction in Ca 2ϩ release (p Ͻ 0.05) (Fig.  2, A and B). Both the pharmacologic and genetic inhibition of the CD38 pathway attenuates the amount of intracellular Ca 2ϩ release observed with TLCS. To examine the nonspecific effects of nicotinamide, Cd38-deficient acinar cells were stimulated with TLCS in the presence or absence of the inhibitor. We observed that nicotinamide caused reductions in the TLCS-induced Ca 2ϩ FIGURE 4. Bile acid-induced cell injury is dependent on CD38. Acinar cells were pretreated with increasing concentrations of nicotinamide for 30 min prior to a 4-h incubation with TLCS (500 M). Cell injury was measured by percent LDH release (A) and propidium iodide uptake (B). C, human acinar cells were obtained from a cadaveric specimen and stimulated with TLCS (500 M) with or without nicotinamide. Cell injury was measured as percent LDH release. Wild-type or Cd38-deficient acinar cells were stimulated with TLCS (500 M) for 6 h, and cell injury was assessed by LDH release (D) and propidium iodide uptake (E) (n ϭ 3). * and #, p Ͻ 0.05 compared with the control and TLCS alone, respectively. RLU, relative light units. signal (Fig. 2, C and D), suggesting that this inhibitor has nonspecific targets that are in addition to inhibiting CD38.
Bile Acid-induced Ca 2ϩ Signals Are Dependent on cADPR-Bile acids target the RyR to induce aberrant acinar cell Ca 2ϩ signals and cell injury (18). Because CD38 drives the synthesis of the RyR activator cADPR, we next asked whether bile acidinduced Ca 2ϩ release is mediated by cADPR.
We found that pretreatment with the cADPR inhibitor 8-Br-cADPR (30 M) reduced the TLCS-stimulated Ca 2ϩ transient by 75% (p Ͻ 0.05) (Fig. 3). The effect was reversible because a return peak was observed after the inhibitor was washed out. These data suggest that the mechanism for CD38-mediated Ca 2ϩ release following TLCS stimulation is through cADPR.
Bile Acid-induced Cell Injury Is Dependent on CD38-cADPR-Bile acids cause injury to pancreatic acinar cells (18,42,43). The injury is dependent on an aberrant rise in cytosolic Ca 2ϩ release. To examine whether CD38 is involved in the pathogenesis of acinar injury, we pretreated isolated acinar cells with varying concentrations of nicotinamide prior to TLCS stimulation. Higher concentrations of nicotinamide abrogated TLCSinduced LDH release (p Ͻ 0.05) (Fig. 4A). In addition, nicotinamide significantly reduced propidium iodide uptake (p Ͻ 0.05) (Fig. 4B).
To determine the relevance of the current findings to the human condition, we obtained live human pancreatic acinar cells from a 9-year-old female donor who died of anoxia. The cold ischemia time was ϳ13 h. The cells were stimulated with TLCS in the presence or absence of nicotinamide. Similar to what was seen in rodent acini, TLCS (500 M) caused an increase in LDH leakage (p Ͻ 0.05) (Fig. 4C). In addition, nicotinamide prevented cell injury (p Ͻ 0.05). Although the data are limited due to the availability of fresh human samples and nonspecific effects of the CD38 inhibitor, they provide relevance to the experimental finding that CD38 plays a critical role in mediating bile acid-induced acinar cell injury.
To further examine the role of CD38 and cADPR in cell injury, we isolated acinar cells from Cd38 Ϫ/Ϫ mice and stimulated them for 6 h with TLCS (500 M). Cell injury measurements revealed that Cd38-deficient acinar cells were protected against TLCS-induced cell injury over the duration of the 6-h time course (Fig. 4, D and E). We also observed reductions in TLCS-induced cell injury (down to the base line) following pretreatment with 8-Br-cADPR (Fig. 5). To understand whether premature activation of digestive enzymes contributes to this injury, we measured the activity of the protease chymotrypsin from wild-type and Cd38-deficient mice simulated with TLCS. Fig. 6 shows that 1 h after stimulation, TLCS induced a 6-fold increase in chymotrypsin activity. The activity was reduced by 20% relative to control levels in Cd38-deficient acinar cells (p Ͻ 0.05). Taken together, these data demonstrate the importance of CD38 and cADPR in bile acid-induced cell injury and protease activation.
Cd38 Ϫ/Ϫ Mice Are Protected against Bile Acid Infusion Pancreatitis-To examine the clinical relevance of CD38-cADPR in the intact animal, we employed an in vivo model of bile acid infusion in which wild-type or Cd38 Ϫ/Ϫ mice received a brief retrograde duct infusion of TLCS (3 mM). We evaluated pancreatic sections for early indices of acute pancreatitis, including edema, inflammation, vacuolization, and necrosis. Remarkably, the Cd38-deficient mice were protected against pancreatitis. Each of the histological parameters of pancreatic injury was reduced to control levels (Fig. 7).

DISCUSSION
The key findings of this study are that CD38 and cADPR mediate TLCS-induced Ca 2ϩ release, acinar cell injury, and bile acid infusion pancreatitis in vivo. Although it is known that bile  Acinar cells from wild-type or Cd38-deficient mice were stimulated with TLCS (500 M). Chymotrypsin activity was measured 1 h later. Data are expressed as -fold increase versus control (n ϭ 3). * and #, p Ͻ 0.05 compared with the control and TLCS alone, respectively. SEPTEMBER 20, 2013 • VOLUME 288 • NUMBER 38 acids trigger several Ca 2ϩ -mediated injurious pathways within the acinar cell, neither the origin of the aberrant Ca 2ϩ signals nor their targets have been fully clarified. Bile acids may work through ligand binding of a G protein-coupled bile acid receptor (Gpbar1; also known as Tgr5) or through bile transport into the cell (7). They can cause sustained Ca 2ϩ release, which then triggers opening of store-operated Ca 2ϩ entry channels (7). In addition, several targets of aberrant acinar cell Ca 2ϩ following bile acid exposure have been suggested and include mitochondria (44,45), sarco/endoplasmic reticulum Ca 2ϩ -ATPase (SERCA) pumps (7), and a host of Ca 2ϩmediated proteins (28,42). We have shown that the serine/ threonine phosphatase calcineurin is activated in response to Ca 2ϩ generated by bile acid exposure (42) and that calcineurin plays an important role in regulating bile acid-induced NF-B activation (46).

CD38 in Pancreatitis
In the acinar cell, bile acids potentiate Ca 2ϩ release from the endoplasmic reticulum and vesicular Ca 2ϩ stores through opening of both inositol 1,4,5-trisphosphate receptors and RyRs (5,28). TLCS-induced Ca 2ϩ transients were reduced by the inositol 1,4,5-trisphosphate receptor inhibitor caffeine (5) and in two-photon permeabilized acinar cells by the RyR inhib-itor ruthenium red (28). The latter result is of particular interest because RyR-dependent Ca 2ϩ release in the pancreatic acinar cell is a potential target for blocking aberrant Ca 2ϩ signals. The RyR was initially shown to mediate pancreatitis in a secretagogue hyperstimulation model of pancreatitis (10,15) and, subsequently, during bile acid exposure (18,28). cADPR was first identified as a critical Ca 2ϩ second messenger in sea urchins (47,48). Okamoto and co-workers (49) demonstrated that cADPR modulates stimulus-secretion coupling in the pancreas using permeabilized mouse islets. Several groups have demonstrated a role for cADPR in the generation of cytosolic Ca 2ϩ signals in the pancreatic acinar cell. First, either intracellular application of cADPR (50) or localized uncaging of cADPR (51,52) can induce Ca 2ϩ signals or cause depletion of the intracellular Ca 2ϩ pool (53)(54)(55). Second, cADPR levels within the acinar cell rise during both acetylcholine and cholecystokinin stimulation (30,56,57). Third, the cADPR antagonist 8-amino-cADPR inhibits Ca 2ϩ transients induced by cholecystokinin (58), acetylcholine (51), or bombesin (59). To our knowledge, this work is the first to show that cADPR also mediates acinar cell Ca 2ϩ signals and injury following bile acid exposure. . Cd38 ؊/؊ mice are protected against bile acid infusion pancreatitis. TLCS (3 mM) was infused into the pancreatic duct of wild-type or Cd38 Ϫ/Ϫ mice. Tissues were collected 24 h after infusion. A, representative H&E sections from the pancreatic head. B and C, overall severity scores and subscores, respectively (n ϭ five to seven animals/group). * and #, p Ͻ 0.05 compared with the wild-type control and TLCS, respectively. cADPR is synthesized from NAD ϩ by ADP-ribosyl cyclases. The main ADP-ribosyl cyclase that generates cADPR in the acinar cell appears to be CD38 because Cd38 Ϫ/Ϫ acini have reduced cADPR levels and fail to increase cADPR levels following secretagogue stimulation (30,56). In contrast, non-CD38 ribosyl cyclases are active in brain (60) and may be contributed by another member of the superfamily of ribosyl cyclases, CD157, also known as BST-1 (61)(62)(63). Several important physiologic roles have been associated with CD38, including the control of insulin secretion (64), clearance of bacterial infections (60), and altered social behavior (65). Our findings using primarily Cd38 Ϫ/Ϫ mice demonstrate that CD38 regulates pancreatic acinar cell injury and in vivo pancreatitis following exposure to bile acids.
Our data raise a few unresolved questions in the CD38 field (47). First, it is unclear how a stimulus, e.g. a bile acid, regulates CD38. The regulation of CD38 by cGMP (66) or cAMP-mediated phosphorylation (67) has been suggested. Second, because CD38 is an ectoenzyme, there is the topological paradox of how a cytosolic substrate such as NAD ϩ comes into contact with the active site of CD38, which faces the extracellular space, and then how its product cADPR is shuttled back into the cell (68). Third, CD38 synthesizes NAADP and NAD ϩ , two other nucleotides besides cADPR that can affect Ca 2ϩ signals. NAADP is derived from NADP and is a potent activator of the two pore channels in endolysosomes (69). The other nucleotide made by CD38 from NAD ϩ is the non-cyclic compound ADPR, which activates the Ca 2ϩ influx channel TRPM2 (transient receptor potential cation channel member 2) (60,70). For this reason, our Ca 2ϩ signaling studies were designed to exclude the influence of Ca 2ϩ influx by using Ca 2ϩ -free medium. Nonetheless, we cannot exclude the possibility that CD38-generated ADPR contributes to acinar cell injury or pancreatitis due to bile acids. In summary, we offer the first demonstration that bile acids cause acinar cell Ca 2ϩ release, injury, and pancreatitis through CD38. These data provide further insight into the Ca 2ϩ -dependent mechanisms of bile acid-induced pathology and suggest that the CD38-cADPR pathway may serve as a target for the treatment of some forms of pancreatitis.