Pancreas-specific SNAP23 depletion prevents pancreatitis by attenuating pathological basolateral exocytosis and formation of trypsin-activating autolysosomes

ABSTRACT

Intrapancreatic trypsin activation by dysregulated macroautophagy/autophagy and pathological exocytosis of zymogen granules (ZGs), along with activation of inhibitor of NFKB/NF-κB kinase (IKK) are necessary early cellular events in pancreatitis. How these three pancreatitis events are linked is unclear. We investigated how SNAP23 orchestrates these events leading to pancreatic acinar injury. SNAP23 depletion was by knockdown (SNAP23-KD) effected by adenovirus-shRNA (Ad-SNAP23-shRNA/mCherry) treatment of rodent and human pancreatic slices and in vivo by infusion into rat pancreatic duct. In vitro pancreatitis induction by supraphysiological cholecystokinin (CCK) or ethanol plus low-dose CCK were used to assess SNAP23-KD effects on exocytosis and autophagy. Pancreatitis stimuli resulted in SNAP23 translocation from its native location at the plasma membrane to autophagosomes, where SNAP23 would bind and regulate STX17 (syntaxin17) SNARE complex-mediated autophagosome-lysosome fusion. This SNAP23 relocation was attributed to IKBKB/IKKβ-mediated SNAP23 phosphorylation at Ser95 Ser120 in rat and Ser120 in human, which was blocked by IKBKB/IKKβ inhibitors, and confirmed by the inability of IKBKB/IKKβ phosphorylation-disabled SNAP23 mutant (Ser95A Ser120A) to bind STX17 SNARE complex. SNAP23-KD impaired the assembly of STX4-driven basolateral exocytotic SNARE complex and STX17-driven SNARE complex, causing respective reduction of basolateral exocytosis of ZGs and autolysosome formation, with consequent reduction in trypsinogen activation in both compartments. Consequently, pancreatic SNAP23-KD rats were protected from caerulein and alcoholic pancreatitis. This study revealed the roles of SNAP23 in mediating pathological basolateral exocytosis and IKBKB/IKKβ’s involvement in autolysosome formation, both where trypsinogen activation would occur to cause pancreatitis. SNAP23 is a strong candidate to target for pancreatitis therapy.

Abbreviations: AL: autolysosome; AP: acute pancreatitis; AV: autophagic vacuole; CCK: cholecystokinin; IKBKB/IKKβ: inhibitor of nuclear factor kappa B kinase subunit beta; SNAP23: synaptosome associated protein 23; SNARE: soluble NSF (N-ethylmaleimide-sensitive factor) attachment protein receptor; STX: syntaxin; TAP: trypsinogen activation peptide; VAMP: vesicle associated membrane protein; ZG: zymogen granule.

Introduction

Acute pancreatitis (AP) is a major health problem that culminates from multi-organelle dysfunction and pathological cellular interactions [1,2]. Postprandial neurohormonal stimulation of pancreatic acini physiologically releases inactive digestive proenzymes stored in zymogen granules (ZGs) by exocytotic fusion exclusively at the apical plasma membrane (PM) [3,4]. Pancreatic acini are vulnerable to stress that rapidly progresses to injury, thus requires an efficient mechanism, which is autophagy, to maintain homeostasis [5–7]. Autophagy sequesters damaged organelles (mitochondria, ZGs) and cytotoxic debris (protein aggregates like excess secretory proteins) and deliver them to lysosomes (autolysosomes) for degradation and recycling [8,9]. AP stimuli (supraphysiological CCK, acetylcholine or ethanol plus low CCK) cause perturbation in both autophagy and exocytosis leading to acinar injury [10,11]. In AP, this perturbation of autophagy is at the autolysosome maturation and clearance steps, which result in inactive trypsinogen becoming prematurely activated into trypsin in autolysosomes by lysosomal cathepsin B [12,13] that then spills into the cytosol to cause injury [14]. Thus, in AP, mild attenuation of autophagy [12] and autolyososme formation [14] could be protective, whereas enhancement of autolysosome formation but with hindrance of its maturation accentuated disease severity [15]. In AP, physiological apical exocytosis of ZGs is blocked and redirected to the basolateral PM to release proteases into the interstitial space [11,16] where subsequent activation by and/or modulation of immune cells can accelerate injury [17,18]. This pathological immune-modulation also involves the inhibitor of NFKB/NF-κB kinase (IKBKB/IKKβ), which is activated in parallel to trypsinogen activation [19,20] to promote NFKB-mediated production of proinflammatory chemokines and cytokines that initiates and perpetuates the inflammatory response [21,22]. IKBKB/IKKβ also regulates autophagy in a NFKB-dependent and -independent fashion [23,24]. We here embarked on searching for a common regulatory component that link these pathological processes, which would provide a common target and thus more effective therapeutic intervention for AP.

Membrane fusion is canonically mediated by soluble N-ethylmaleimide-sensitive-factor attachment receptor (SNARE) proteins by specific pairing of isoforms of target membrane SNAREs (t-SNAREs: syntaxins [STXs] and SNAP25 [synaptosome associated protein 25]) and vesicle-associated SNAREs (v-SNAREs: vesicle-associated membrane proteins or VAMPs) [25]. These membrane fusion events include secretory vesicle exocytosis and numerous intracellular vesicle trafficking [25], including autophagy [26]. SNAREs were found to also interact with non-SNARE partners, as is the case in autophagy. In autophagy, distinct SNARE proteins regulate the biogenesis of phagophores (STX7, VAMP7, VTI1B) [27], autophagosomes (STX17) [28,29] and autophagosome fusion with lysosomes (STX17, YKT6, SNAP29, VAMP8) [30,31]. Our recent work has identified STX2 as a regulator of acinar phagophore biogenesis [15].

Ubiquitously expressed SNAP23 mediates acinar secretion through its involvement in physiological apical and pathological basolateral exocytosis of ZGs [4,32–34]. Genetic deletion of SNAP23 in mouse pancreatic acini confirmed its role in apical exocytosis [35]. SNAP23’s role in autophagy was recently suggested in a report on adipocyte-specific SNAP23 deletion [36]. SNAP23 was found to be a substrate of IKBKB/IKKβ [37], the latter well known to be highly activated in AP [19,21,22]. Taken together, this raises the possibility that SNAP23 could be the common link in these apparently divergent pathological cellular processes in AP. We here advanced the hypothesis for a role for SNAP23 in autolysosome formation by its action on the STX17 autophagosome-lysosome fusogenic SNARE complex, which involves IKBKB/IKKβ activation of SNAP23. We show that pancreas-specific SNAP23 depletion (SNAP23-KD) abrogated all these events, and also basolateral exocytosis, which together contributed to a major reduction in disease severity of AP.

Results

SNAP23-KD reduces cholecystokinin (CCK)-stimulated amylase secretion and perturbs pancreatitis induced autophagy

To reduce endogenous levels of SNAP23, we used the strategy of Ad-SNAP23-shRNA/mCherry with shRNAs targeted at the rat (CCGGGAGTCTGGCAAGGCTTATAAGCTCGAGCTTATAAGCCTTGCCAGACTCTTTTTG) and human (CCGGGAACAACTAAATCGCATAGAACTCGAGTTCTATGCGATTTAGTTGTTCTTTTTG) SNAP23 mRNAs. We empirically determined (Figure S1A,B) the volume of virus suspensions (1.0 uL/mL for rat; 1.5 uL/mL for human) that infected >90% of the pancreatic slice area to achieve knockdown (KD) efficiency of ~70% in rat (Figure 1A) and ~66% in human (Figure 1B) pancreatic slices over 48 h, which did not affect the expression of other exocytotic SNAREs and SNAP25 isoform SNAP29 (Figure S1C,D). Ad-sc-shRNA/RFP virus-infected tissues were used as ῝Control”, marked WT.

Figure 1.

SNAP23 depletion in acini inhibits amylase secretion and perturbs pancreatitis-induced autophagy. (A,B) Rat (A) and human (B) pancreas slices were treated with Ad-SNAP23-shRNA/mCherry (SNAP23-KD/KD) vs Ad-sc(scramble)-shRNA/RFP (control/WT) for 48 h. Western blot analysis (left) and densitometry analyses (right, N = 3, normalized to loading control TUBA4A/tubulin) showed reduced SNAP23 expression (see Results). (C-F) SNAP23-KD reduced AMY/amylase secretion in rat (C,D) and human (E,F) pancreatic slices as assessed by a colorimetric assay (C,E; secreted AMY/amylase expressed as percentage of total pancreatic slice AMY/amylase, n = 8 from 3 independent experiments); and as detected by western blots (D,F; left) of AMY/amylase released into the buffer or retained in the slices, expressed as a densitometry ratio (D,F; right) of WT (100 pM for buffer and 0 pM for slices) as 100. (G,H) Western blot detection of autophagic vacuole (AV) marker protein LC3B-II and autophagy substrate protein SQSTM1/p62 (left- representative western blots; right- densitometry analyses (N = 3) by the ratio of LC3B-II or SQSTM1/p62 to loading control ACTB/actin) in human pancreatic slices exposed to pancreatitis stimuli (G: 10 nM CCK; H: 100 pM CCK plus 20 mM EtOH) vs starvation (G). (I) Increased accumulation of GFP-LC3B (by Ad-GFP-LC3B co-infection) puncta (which are AVs) in SNAP23-KD human acini (left) induced by pancreatitis stimuli. Quantification of GFP-LC3B puncta is shown on right from 30 acinar cells from 3 independent experiments. AMY/amylase secretion, densitometry and quantification values are expressed as mean±S.E.M; NS = Not significant. *, P< 0.05

Both rat and human pancreatic slices displayed classical biphasic pattern of amylase secretion (Figure 1C,E) similar to dispersed acini secretory response [3,4]. SNAP23-KD effected an inhibition of secretion stimulated by submaximal (50 pM, by 42.6% in rat) and maximal (100 pM, by 35% in rat, 32% in human) CCK, known to effect apical exocytosis; and also supramaximal CCK (10 nM, by 47.5% in rat, 41% in human) inhibitory response, known to block apical exocytosis with redirection of ZG fusion to the basolateral PM [4,38]. To assess whether the reduced amylase secretion from SNAP23-KD slices might be attributed to inadvertent reduction in zymogen synthesis, we confirmed by immunoblotting the reduction in levels of amylase in the buffer after 100 pM CCK (reduced by 32% in rat and 28% in human) and 10 nM CCK (reduced by 42% in rat and 37% in human) stimulation; and importantly, there were small but insignificant increases in amylase retention in the SNAP23-KD slices (Figure 1D,F) [39]. The sum of the densitometric values of the cellular and secreted amylase (data not shown) showed equivalent amylase content in WT and SNAP23-KD slices. Therefore, the reduction in amylase secretion from SNAP23-KD pancreas slices (Figure 1C-F) was a result of defects in secretion and not from a change in acinar amylase content, indicating that SNAP23-KD does not affect zymogen synthesis. To assess the impact of SNAP23-KD on autophagy in human pancreatic acini (SNAP23-KD vs WT), we determined the expression levels of lipidated LC3B-II as a marker of autophagic vacuole (AV) formation and the selective autophagy substrate SQSTM1/p62 [40–42] under pancreatitis conditions (stimulation with 10 nM CCK (Figure 1G) or 20 mM EtOH + 100 pM CCK (Figure 1H)) compared to physiological autophagy induction by starvation (in EBSS for 2 h). Both starvation and pancreatitis stimuli augmented the accumulation of LCB-II. Whereas SNAP23-KD had no effects on starvation autophagy induction of LC3B-II, SNAP23-KD further increased the LC3B-II levels in pancreatitis autophagy induction, by 88% after 10 nM CCK (Figure 1G) and 84% after 20 mM EtOH + 100 pM CCK (Figure 1H) stimulation. SQSTM1/p62 levels were depleted to similar extent in SNAP23-KD vs WT acini on starvation but displayed a 208% (Figure 1G) and 72% (Figure 1H) further elevated accumulation in SNAP23-KD compared to WT acini. Actual visualization of LC3B-positive puncta [40,41] (images in Figure 1I) showed 89% increase in AVs in SNAP23-KD acini after 10 nM CCK pancreatitis stimulation but no change in the starvation condition (data not shown). These results showing that SNAP23-KD could cause perturbation in amylase secretion and pancreatitis-related autophagy motivated us to further assess the SNAP23-KD effects on ZG exocytosis (Figure 2) and the autophagy events occurring in pancreatitis (Figure 3).

Figure 2.

SNAP23-KD of human acini reduces pathological basolateral exocytosis by impairing the formation of its exocytotic SNARE complexes. (A) Analysis of exocytosis. 36 h post-infected WT and SNAP23-KD pancreatic slices were co-infected with Ad-syncollin-pHluorin for another 12 h, then dispersed into acini and plated on coverslips for imaging. Acini were subjected to pancreatitis (10 nM CCK-8) stimulation. Representative images from the indicated time points are shown in left. ZG fusion events observed as pHluorin hotspots were assigned to “apical” and “basolateral” areas [15,43,44,57], and normalized to cell area and recording time (right). Data (N = 6) expressed as mean±SEM, *P< 0.05, Scale bars: 10 μm. (B,C) Analysis of STX3 (B) and STX4 (C) exocytotic SNARE complex formation in WT vs SNAP23-KD pancreatic slices after pancreatitis (10 nM CCK-8) stimulation. Degree of association are expressed as band intensity ratios of indicated co-IPed proteins that immunoprecipitated (IP) with STX3 (B) or STX4 (C) in right and expressed as mean±SEM (N = 3). Set-1 and Set-2 are 2 independent sets of experiments. Red asterisks showing reduced levels of co-IPed VAMP8 in SNAP23-KD pancreas. “Input” controls (10% or 20 μg total acini lysates) with densitometric analyses are shown in Figure S2A,B

Figure 3.

SNAP23-KD in human acini attenuates pancreatitis-induced autolysosome formation by blocking autophagosome-lysosome fusion. (A) WT and SNAP23-KD pancreatic slices were kept under starvation condition (Starv; 2 h in EBSS) or treated with Baf A1 (2 h), or subjected to pancreatitis stimulus (10 nM CCK-8, 1 h) alone or in presence of salubrinol, Baf A1 or 3-MA. Level of LC3B-I conversion to LC3B-II was assessed by western (top) and quantified and expressed as densitometric ratio to loading control ACTB/actin (bottom). (B) WT and SNAP23-KD slices were cultured as in Figure 2 and co-infected with Ad-GFP-mRFP-LC3B for 12 h, then dispersed into acini that were plated and fixed immediately as Control (top panel), or exposed to starvation condition (middle panel) or stimulation with 10 nM CCK-8 (bottom panel), then fixed and imaged. Yellow puncta indicate autophagosomes, red puncta indicate autolysosomes. Scale bars: 10 µm. (C) Analyses of the two AV populations (autophagosomes: yellow bars and autolysosomes: red bars) from 42 WT and 61 KD acinar cells (at least 10 cells from each experiment) from 3 independent experiments. (D) Immunofluorescence images (enlarged) of transiently expressed GFP-LC3B with native LAMP1 (lysosome marker protein) in Control (top panel), or 10 nM CCK-8-stimulated (lower panel) WT and KD acini. Scale bars: 10 µm. Complete sequences of images demarcated with enlarged regions are shown in Figure S3. Quantification of autolysosomes (colocalized yellow puncta) from 17 acini (WT) and 22 acini (KD) from 2 independent experiments is shown at the bottom. (E,F) Defect in CTSL (E) and CTSB (F) processing and their activities during 10 nM CCK-8 pancreatitis stimulation were similar in the lysosomes from WT and SNAP23-KD pancreatic slices. I: intermediate form and SC: single-chain form. (G) Representative images showing increased appearance of TAP-positive puncta in WT acini during pancreatitis stimulation compared to SNAP23-KD acini. (H) Quantification of TAP-positive puncta from experiments related to Figure G. Control (unstimulated) acini displayed very low numbers of puncta (data not shown). (I) Detection of total cellular trypsin activity in WT vs SNAP23-KD acini. Scale bars: 10 μm. Data shown as mean±S.E.M. *, P < 0.05; NS, not significant

SNAP23-KD reduces pathological basolateral exocytosis by impairing its SNARE complex formation

We used Ad-syncollin-pHluorin-treated (12 h) WT and SNAP23-KD pancreatic slices to track exocytosis by real-time imaging [15,43]. Syncollin is a ZG content protein, and pHluorin is a GFP variant that has low fluorescence in acidic pH (intact ZGs) and becomes highly fluorescent appearing as hotspots upon ZG fusion with PM, when pHluorin is exposed to extracellular basic pH. Ad-syncollin-pHluorin-infected slices were dispersed into acini (more amenable for microscopy), which were plated on poly-L-lysine-coated coverslips and imaged immediately under a spinning-disk confocal microscope. As similarly reported by Kunii et al using a SNAP23 knockout mouse [35], we confirmed that apical exocytosis (hotspots at the apical region) stimulated by physiological 10 pM CCK was severely abrogated by the SNAP23 deletion (data not shown, not the focus of this paper). We instead focused on basolateral exocytosis, a basis of pancreatitis [11,15,16,34,44] that was not investigated in the Kunii report. In Figure 2A, the inner circle was drawn over the apical pole, distinguished from the basaolateral region encompassed between the inner and outer circles. With 10 nM CCK pancreatitis stimulation, apical exocytosis hotspots were partially blocked with abundant exocytosis occurring at the basolateral plasma membrane (Figure 2A) as previously reported [15,34,44]. After SNAP23-KD, there was major reduction in 10 nM CCK-stimulated basolateral exocytosis hotspots as well as further reduction in the already abrogated apical exocytosis, contributing to the greater reduction in amylase secretion (Figure 1C-F). Consistently, analysis of the exocytotic SNARE complexes by immunoprecipitation (IP) showed that SNAP23-KD caused a perturbation in the assembly of both the apical (STX3-SNAP23-VAMP2-VAMP8) (Figure 2B, indicated by red asterisk) and basolateral (STX4-SNAP23-VAMP8) [15,33,34] (Figure 2C, indicated by red asterisks) SNARE complexes. Much work has previously been reported to characterize these apical and basolateral exocytotic SNARE complexes [15,33,34].

SNAP23-KD attenuates pancreatitis-induced pathological autolysosome formation by blocking autophagosome-lysosome fusion

To elucidate the underlying mechanism by which SNAP23-KD alters autophagy (Figure 1G-I), we biochemically analyzed the autophagic flux [40,41] quantitated by the amounts of lipidated LC3B-II. Starvation induction of autophagy showed no discernible effect (Figure 3A), whereas 10 nM CCK pancreatitis stimulation increased LC3B-II accumulation, which was further increased by SNAP23-KD (Figure 3A). Since SNAP23-KD impairment of CCK-stimulated secretion (Figure 1C-F) might involve activation of ER-stress or ER-stress-induction of autophagy [44], we used an ER-stress-induced autophagy inhibitor salubrinol [44] which partially restored LC3B-II conversion (Figure 3A), indicating that the increased AV accumulation caused by SNAP23-KD was attributed to additional factor(s). Furthermore, pre-treatment with bafilomycin A1 that blocks autophagosome-lysosome fusion and/or lysosomal degradation resulted in an equivalent increase in LC3B-II accumulation in WT and SNAP23-KD acini either in the presence or absence of 10 nM CCK stimulation; whereas an autophagy initiation inhibitor 3-methyl adenine (3-MA) reduced LC3B-II to a similar extent (Figure 3A). These results indicate that the increased AV accumulation caused by SNAP23-KD during pancreatitis stimulation is attributed to a perturbation in a step upstream of autolysosome formation and maturation.

We therefore distinguished (Figure 3B) and quantified (Figure 3C) autophagosomes from autolysosomes using GFP-mRFP-LC3B- imaging [40] whereby autophagosomes at neutral pH fluoresce yellow, whereas autolysosomes at acidic pH fluoresce red since the acidic pH quenches GFP fluorescence. At both basal (Figure 3B, top panel and Figure 3C) and starvation (Figure 3B, middle panel, analysis in Figure 3C) conditions, there were equal numbers of autophagosomes (yellow puncta) between WT vs SNAP23-KD acini. With 10 nM CCK pancreatitis stimulation (Figure 3B, bottom panel and Figure 3C), WT acini displayed greatly increased accumulation of autolysosomes (red hotspots), consistent with earlier reports [13,15]. However, pancreatitis stimulation of SNAP23-KD acini instead showed autophagosomes greatly outnumbering autolysosomes (more yellow puncta than red puncta in Figure 3B, bottom panel and Figure 3C), consistent with the increase in LC3B-positive puncta in Figure 1I. When comparing only the autolysosomes, SNAP23-KD acini showed a 74% reduction (Figure 3C). To ascertain that this was due to reduction in autolysosome formation caused by SNAP23-KD, we more critically assessed in CCK-stimulated SNAP23-KD acini for the formation of autolysosomes by quantifying the colocalization (yellow) of transiently expressed GFP-LC3B with endogenous LAMP1, a lysosome marker protein (red) in Figure 3D. Complementary to our Ad-GFP-mRFP-LC3B data (Figure 3B,C), we observed increased number of AVs (green puncta) in SNAP23 KD acini. However, our analysis disclosed a significant 56% dip in autolysosome numbers (yellow puncta, Figure 3D) in CCK-stimulated SNAP23-KD cells compared to WT cells (WT: 70.64% of total AVs are autolysosomes; KD: 31.12% of total AVs are autolysosomes). Lysosomal CTSL (cathepsin L)-CTSB processing and activities remained similarly defective (CTSL accumulated intermediate (I) and CTSB accumulated single chain (SC) isoforms to similar extent) [13] in WT and SNAP23-KD acini (Figure 3E,F) upon 10 nM CCK stimulation, indicating that SNAP23-KD had no effect on the pancreatitis perturbation of this lysosome maturation step [13]. To further investigate autolysosome maturation, we assessed for pathological trypsin activation known to occur in pancreatitis-affected autolysosomes [12–15], using trypsin activation peptide (TAP) staining (Figure 3G). In agreement with the reduced number of autolysosomes (Figure 3B-D), SNAP23-KD acini showed 63% less TAP-positive puncta (Figure 3G,H), along with reduction of total cellular trypsin activity by 52% (Figure 3I). Taken together, SNAP23-KD specifically blocked autophagosome fusion with lysosomes, but those autolysosomes that were able to form, SNAP23-KD did not affect the downstream steps of autolysosome maturation, including cathepsin processing and trypsin activation, which remained equally perturbed (SNAP23-KD vs WT) by the pancreatitis stimulation.

Pancreatitis stimuli promote SNAP23 association with the STX17 SNARE complex required for autolysosome formation, which is disabled by SNAP23-KD

SNAP23-KD-induced blockade of autophagosome-lysosome fusion suggests an effect on the putative SNARE complex(es) that mediate fusion of these two vesicles, including STX17 on autophagosomes that forms a complex with SNAP29 and lysosomal VAMP8 [30], and STX7 on lysosome, which forms a complex with SNAP29 and YKT6 on autophagosomes [31]. Pancreatitis stimulation (10 nM CCK or 100 pM CCK+20 mM EtOH) of human pancreas slices increased STX17 IP pulldown of its SNARE complex (Figure 4A), which was remarkably disrupted by SNAP23-KD, shown as reduced assembly of STX17 with lysosomal VAMP8 without affecting the binding to SNAP29. The STX17 SNARE complex also bound SNAP23 (Figure 4A). The alternative STX7 SNARE complex was also induced to increase by the pancreatitis stimuli (Figure 4B), but this complex did not bind SNAP23 and was not affected by the SNAP23-KD. In contrast, starvation, while also able to induce formation of both STX17 and STX7 SNARE complexes, both these complexes did not bind SNAP23, and therefore were not disrupted by the SNAP23-KD (Figure 4C,D). These results indicate that pancreatitis-induced pathological excess of autophagosome-lysosome fusion requires SNAP23 to associate with and stabilize the STX17 SNARE complex, particularly with lysosome VAMP8; whereas this SNAP23 association is not required for STX17 or STX7 SNARE complexes in physiological starvation that mediates normal autolysosome formation.

Figure 4.

Pancreatitis stimuli promote SNAP23 association with the STX17 SNARE complex required for autolysosome formation, which is disabled by SNAP23-KD. (A-D) Assessment of autophagosome-lysosome fusion SNARE complexes (STX17-SNAP29-VAMP8 and STX7-SNAP29-YKT6) were analyzed on WT and SNAP23-KD human pancreatic slices by co-IP. Pancreatitis stimulation (10 nM CCK-8 or ethanol pancreatitis stimulation) of human pancreatic slices induced SNAP23 association with STX17-SNAP29-VAMP8 complex (A) but not with STX7-SNAP29-VAMP8 complex (B); and for both not in slices exposed to starvation condition (C,D). STX17 co-IPed less amounts of VAMP8 from SNAP23-KD slices (A) whereas SNAP23-KD had no effect on STX7-SNAP29-YKT6 complex formation (B). Association of proteins are expressed in the corresponding right panels as band intensity ratios of indicated co-IPed proteins to IPed STX17 (A,C) or STX7 (B,D). “Input” controls and their densitometric analyses are shown in Figure S4. Data shown as mean±S.E.M (N = 3). *, P < 0.05; NS, not significant

SNAP23 translocation from the plasma membrane to STX17-containing autophagosomes is mediated by IKBKB/IKKβ -phosphorylation of SNAP23

We next examined how SNAP23, normally localized and most abundant in PM [32,34,35], could associate with the STX17 SNARE complex expected to occur on the AV. At basal condition, we confirmed SNAP23 abundance in the acinar cell PM (red on PM, Figure 5A left image and Figure S5A, left image). With 10 nM CCK-8 pancreatitis stimulation, PM-bound SNAP23 was depleted, and instead was displaced into the cytoplasm that became co-localized in STX17-positive intracellular vesicles (green puncta), appearing as yellow puncta (Figure S5A, right image). Quantification showed a 66% increase in SNAP23-positive STX17 puncta (- CCK: 1.27% of total STX17 puncta, + CCK-8: 67.25% of total STX17 puncta) on pancreatitis stimulation (Figure S5B). Using Ad-GFP-LC3B-infected acini subjected to pancreatitis stimuli (10 nM CCK or 100 pM CCK+20 mM EtOH), we confirmed that the SNAP23 translocation to STX17 vesicles were in fact autophagosomes (Figure 5A, middle and right images). Specifically, our analysis showed increases of 61% in 10 nM CCK-stimulated acini and 47% in 100 pM CCK+20 mM EtOH-stimulated acini in SNAP23 localization to STX17-positive autophagosomes (Figure 5B).

Figure 5.

IKBKB/IKKβ-mediated phosphorylation of SNAP23 is required for SNAP23 to translocate to STX17-containing autophagosomes. (A) Representative images showing translocation of SNAP23 from PM to intracellular STX17-positive GFP-LC3B-puncta (autophagosomes) after pancreatitis stimulation (middle: 10 nM CCK-8; right: ethanol pancreatitis stimulation). Individual channels are shown in magnified boxed regions at the bottom. Scale bars: 10 µm. (B) Quantification of SNAP23-positive STX17-containing autophagosomes. 30 acinar cells were quantified from 3 independent experiments. (C-F) Rat (C,D) and human (E,F) pancreatic SNAP23 are a substrate of IKBKB/IKKβ-induced phosphorylation during pancreatitis stimulation. Quantifications of phosphorylation in the specified conditions are shown as densitometric ratio of phosphorylated SNAP23 (p-Ser95 in C; p-Ser120 in D-F) to total SNAP23 in the corresponding right panels. TUBA4A/tubulin is the loading control. (G) IKBKB/IKKβ inhibition prevented the translocation of SNAP23 to the autophagosomes (- CCK-8, + CCK-8, scale bar: 10 µm; + CCK-8/IKK-In, scale bar: 9 µm). (H) Quantification of GFP-LC3B puncta associated with SNAP23. (I) Quantification of GFP-LC3B puncta in the indicated conditions. 25 acinar cells were analyzed from 3 independent experiments for H and I. (J) Immunoprecipitation assessment of STX17 association with overexpressed WT vs IKBKB/IKKβ-phosphorylation-disabled SNAP23 in AR42J cells after pancreatitis stimuli. Co-IPed proteins within the complexes were analyzed as in the IP analyses in Figures 2 and Figures 4, and expressed as mean±S.E.M at the bottom. Input controls and their densitometry analyses are in Figure S5D. *, P < 0.05; NS, not significant

We investigated the signaling pathway that accounts for the SNAP23 translocation to STX17-containing autophagosomes. In mast cells, SNAP23 translocated from PM to intracellular vesicles during degranulation-stimulation [45], and was found to be a substrate of IKBKB/IKKβ in mast cells and platelets [37,46]. We confirmed that pancreatitis stimuli could induce IKBKB/IKKβ-mediated phosphorylation of SNAP23 in rat (at Ser95 and Ser120, Figure 5C,D) and human (at Ser120, Figure 5E,F) pancreatic acini by using the same specific anti-phospho-Ser95 and anti-phospho-Ser120 antibodies used in the mast cell and platelet studies [37,46]. This was confirmed by the blockade of SNAP23 phosphorylation by specific IKBKB/IKKβ inhibitors BMS-345541 and BAY 11–7082 [23,46] (Figure 5C-F). Pretreatment of acini with BMS-345541 (IKK-In) prior to pancreatitis stimuli resulted in retention of SNAP23 on the PM, preventing its translocation to intracellular vesicles (Figure 5G,H), and a consequent 88.5% reduction (+ CCK: 67.7%; + CCK/+ IKK-In: 7.8%) in SNAP23-positive AVs (Figure 5H). Consistent with a major postulated pleiotropic action of IKBKB/IKKβ in autophagy regulation [23,24], which is IKBKB/IKKβ inhibition-induced suppression of autophagy [23], we also found in IKK-In pretreated acini a modest 38% reduction in the number of AVs (+ CCK/- IKK-In: 2.33 AVs/acinus; + CCK/+ IKK-In: 1.45 AVs/acinus, Figure 5I). To unequivocally demonstrate that IKBKB/IKKβ-mediated SNAP23 phosphorylation is required for its translocation to and subsequent association with autophagosome-bound STX17, and definitively rule out other possible phosphorylation-independent events, we used the AR42J cells (rat acinar cell line) to overexpress WT or IIKβ phosphorylation-disabled mutant of SNAP23 (Ser95A Ser120A) [37,47]. We first verified the level of overexpression and sensitivity of these SNAP23 proteins to pancreatitis stimuli-induced phosphorylation (Figure S5C). STX17 was IPed to compare its association with WT and mutant SNAP23 along with endogenous SNAP29 (Figure 5J). We obtained abundant association of WT SNAP23 with the STX17-SNAP29 complex whereas the IKBKB/IKKβ-phosphorylation-disabled mutant failed to bind this complex (Figure 5J). In contrast, starvation could not induce IKBKB/IKKβ-mediated phosphorylation of SNAP23 (data not shown). These results unequivocally demonstrate that in AP, acinar SNAP23 is a substrate of IKBKB/IKKβ-phosphorylation-mediated translocation to autophagosomes, a step required to associate with the STX17-SNARE complex to effect the pathologically excessive fusion of autophagosomes with lysosomes. These pancreatitis-stimulated pathological SNAP23 events are dispensable in physiological starvation-induced normal autolysosome formation.

SNAP23 depletion protects rats from pancreatitis

The above in vitro results showing that SNAP23 depletion reduced pancreatitis stimulation- induced basolateral ZG exocytosis (Figure 2A) and pathological excess in autolysosome formation (Figure 3B-D), both locations where premature trypsinogen activation would occur [10], should predict a protective effect of SNAP23-KD on reducing pancreatitis severity. To reduce pancreatic SNAP23 levels in vivo, we infused Ad-SNAP23-shRNA/mCherry (KD rats) directly into the rat pancreatic duct [48,49], to preferentially fill the head (verified by co-infused India Ink, Figure S6A,B). Ad-sc(scramble)-shRNA/RFP virus was infused similarly to serve as Control (WT) rats. After virus treatment, these rats were given a 3-week full recovery to negate any contribution of surgery to the pancreatitis. At 3 weeks post-surgery, SNAP23 expression in the Ad-SNAP23-shRNA/mCherry-infected pancreatic head portion was depleted by ~75%, compared to ~50% depletion in the body and normal levels in the tail, whereas SNAP23 expression in WT (Ad-sc(scramble)-shRNA/RFP-infected) pancreas was equal for head, body and tail (Figure 6A). We subjected these WT (Figure S6C) and pancreatic SNAP23-KD (Figure S6D) rats to AP protocols, including supramaximal caerulein, a CCK analog (Figure 6B-E), and alcoholic pancreatitis (Figure 6F-I). WT pancreas displayed uniform and heightened swelling throughout (Figure S6C). In SNAP23-KD pancreas, there was reduced swelling in the head (note persistent India Ink stain) and body portions (less India Ink stain) compared to the tail portion (no India Ink stain, Figure S6D). If the surgical pancreatic duct infusion had any residual effect, it would have been the head of the pancreas that’s most susceptible to pancreatitis, which is the opposite in the SNAP23-KD pancreas. We also assessed the pancreatic tissue levels of SNAP23 to see whether a reduced SNAP23 level per se might be contributing to undefined processes that could also contribute to the acinar injury. Consistent with our in vitro result (Figure S2B,S4B), we did not see any apparent change in the pancreatic tissue levels of SNAP23 protein expression after in vivo caerulein pancreatitis compared to Control pancreas of saline-treated rats (Figure S6E,F). Histological analysis of acinar injury based on scoring of four criteria (edema, inflammation, necrosis, vacuolization) individually and collectively on a scale of 1–4 [15,44] showed the highest protection in the head portion (looked normal) of SNAP23-KD pancreas, followed by body, compared with maximal injury in the tail (Figure 6C- caerulien pancreatitis; Figure 6G- alcoholic pancreatitis). In WT pancreas, pancreatitis scores were high and comparable between the head, body and tail of the pancreas; and were also comparably high between the tail portions of SNAP23-KD and WT pancreases (Figure 6C,G). The acinar injury was minimal and equivalent between WT and SNAP23-KD pancreases from the saline (Figure S6G-I) and only-ethanol treatments (Figure S6J-L). Consistent with the relative levels of SNAP23 depletion in the SNAP23-KD pancreas, pancreatic tissue trypsin activity was also lower in the head and body portions compared to high levels in the tail portions, the latter comparable to the equally high trypsin activities in head, body and tail portions of WT pancreas (Figure 6D,H; left panels). Tissue myeloperoxidase (MPO) activity (Figure 6D,H; right panels) and corresponding polymorphonuclear leukocyte infiltration scores (inflammatory cells, Figure 6C,G) were lowest in the head of SNAP23-KD pancreas, followed by the body, and unaffected in the tail, the latter also showing similarly high MPO activity and inflammatory cell infiltration as all three portions of WT pancreas. Note that the more sensitive biochemical analysis of whole tissue trypsin and MPO activities were much lower in the SNAP23-KD head portions but were nevertheless only mildly higher than saline or only-ethanol control treatments (performed as control for all pancreas portions) likely because not 100% of the pancreatic tissue in the head was depleted of SNAP23 (Figure 6A). Finally, we measured the clinically-used pancreatitis parameters of circulating pancreatic enzymes amylase (left; Figure 6E,I) and lipase (right; Figure 6E,I). Both serum pancreatic enzymes were lower in pancreatic SNAP23-KD rats, which was remarkable since only the head and some of the body that were depleted of SNAP23 had reduced pancreatitis. These results collectively indicate that SNAP23-KD in pancreas greatly reduced the severity of pancreatitis.

Figure 6.

In vivo SNAP23 depletion of the rat pancreas can provide protection against acute pancreatitis. (A) Representative western blots showing pancreatic duct infusion of Ad-SNAP23-shRNA-mCherry (KD) can reduce SNAP23 expression in rat pancreas, preferentially in the pancreatic head (N = 8), less so in the pancreatic body, and leaving unaltered expression of SNAP23 in the pancreatic tail; the latter having similar level of SNAP23 as those found in all three portions of the pancreatic head, body and tail of Ad-sc(scramble)-shRNA/RFP-infused WT (control) pancreas (N = 8). Densitometry analyses (bottom, normalized to loading control TUBA4A/tubulin) showing effective SNAP23-KD in the head (H) > body (B) >tail (T), but equal in all three portions in WT control. (B,F) Representative H&E-stained images (14X magnification) of WT (top panels, n = 3) vs KD (bottom panels, n = 3) pancreatic “H”, “B” and “T” sections (left to right) after caerulein pancreatitis (B) and alcoholic pancreatitis (F), scale bar: 200 μm. (C,G) Individual (to the left of red dotted line) and combined (Total, to the right of red dotted line) histology scores of acinar injury: edema (Ed), inflammatory cell infiltration (Inf), necrosis (Nec) and vacuolization (Vac), on a scale of 0–4 from “H” (top), “B” (middle) and “T” (bottom) portions of pancreases of caerulein (C) and ethanol (G) pancreatitis-induced rats. H&E images of pancreases of controls (n = 2 each) that were injected saline or only-ethanol, and their histology injury score analyses are shown in Figure S6G-I (for saline) and Figure S6J-L (for only-ethanol). (D,H,E,I) Activities of pancreatic tissue trypsin (left; D,H), MPO (right; D,H), serum amylase (left; E,I) and lipase (right; E,I) for caerulein (D,E) and ethanol (H,I)-induced pancreatitis. Data shown as mean±S.E.M. *, P < 0.05; NS, not significant

Discussion

In AP, the cardinal features include the mixture of ZG proteases with trypsinogen-activating lysosomal cathepsin B within autolysosomes [12–14], infiltration of inflammatory cells into interstitial spaces which activates the ZG proteases released from the pathological basolateral exocytosis [11,17,18], and activation of IKBKB/IKKβ that accentuates the inflammatory response [19–22]. We show in this study that SNAP23 links these three AP cell injury events whereby pancreatitis stimulation induces SNAP23 participation in the STX4-mediated pathological basolateral exocytosis followed by IKBKB/IKKβ phosphorylation of SNAP23 to induce its translocation from its original residence on the basolateral PM to autophagosomes to bind STX17 in mediating excessive autolysosome formation.

We [32] and others [33,35] had previously suggested that SNAP23, as part of ZG exocytotic SNARE complexes, mediate apical exocytosis and basolateral exocytosis, which herein was better shown whereby SNAP23-KD clearly reduced basolateral and also confirmed the blockade of apical exocytosis [35]. We now also showed how a distinct SNARE complex is involved in pancreatitis-induced autolysosome formation [12,13]. Autophagosome formation begins with the phagophore or isolation membrane that expands with incorporation of membranes from various intracellular sources, then the phagophore termini seal to form mature autophagosomes [50]; these processes are mediated by not only autophagy-related (Atg) but also SNARE proteins [26]. Nascent autophagosomes acquire STX17 that readily forms complex with SNAP29 and lysosomal VAMP8 to mediate fusion of these two vesicles, generating autolysosomes [30]. The first clue we observed for a new role of SNAP23 in autolysosome formation was that SNAP23-KD in acini caused an accumulation of autophagosomes (thus autophagosome biogenesis per se is not perturbed) but with much reduced number of autolysosomes, the latter indicating the defect to be in autophagosome fusion with lysosome per se. Pancreatitis stimulation therefore pathologically induces SNAP23’s involvement in the perturbation in autolysosome formation. Pancreatitis stimuli also induces abnormal cathepsin processing, indicating lysosome dysfunction [13], and consequent premature trypsinogen activation in autolysosomes, further indicating defective autolysosome maturation; these pancreatitis-induced defects in the lysosome and autolysosome were not affected by SNAP23-KD. Taken together, these results indicate that SNAP23’s role must be on specifically perturbing autophagosome-lysosome fusion by its actions on its SNARE fusion complex. Physiological starvation however retained the ability to induce formation of autolysosome SNARE complexes, but which did not bind SNAP23 and the SNARE complex assembly was not affected by SNAP23-KD. This may explain why SNAP23 was not previously detected to associate with other autophagy-related SNAREs [26]. It seems that these SNAP23-KD-induced defects in pancreatic acinar cells are very different from those recently reported in adipocyte [36], wherein SNAP23 genetic deletion caused a reduction in autophagosome biogenesis and increased adipocyte cell death resulting in generalized lipodystrophy. That report is in contrast to our study showing SNAP23 depletion’s actions on reducing autolysosome formation (not autophagosome biogenesis), and the reduction in acinar injury and pancreatitis severity.

SNAP23 is anchored to PM through its palmitoylated cysteine residues located in the spacer region [25] and could relocate [45] and stably associate with intracellular compartments that are enriched with phosphatidylethanolamine (PE) [51]. Our study showed that pancreatitis stimuli could induce SNAP23 to relocate from PM to autophagosomes to participate in STX17 SNARE complex formation to mediate autophagosome-lysosome fusion (Figure 5A). SNAP23 in pancreatic acinar cells thus display multiple functions depending on which SNARE complex it forms with, which are SNAP23-STX3 SNARE complex (with VAMP8) that mediates granule-granule fusion in apical exocytosis [34] and SNAP23-STX4 (with VAMP8) that mediates basolateral exocytosis [34]. These SNAP23-STX SNARE complexes do not mediate autophagosome-lysosome fusion. It is SNAP23 that forms a complex with STX17 and VAMP8 that mediates autophagosome-lysosome fusion. So here, it is the STX isoform that dictates the specific compartment for fusion, whereas SNAP23 and VAMP8 are common to the three compartments. SNAP29 too promiscuously associates with numerous STXs to mediate various intracellular fusion events [52], including STX17 and STX7 SNARE complexes shown in our study. The requirement of SNAP23 in pathological and excess autolysosome formation could be partly attributed to SNAP23’s superior stability in binding PE-enriched autophagosomes compared to SNAP29, which lacks cysteine in its spacer region [53]. SNAP29 is also a competitor of α-SNAP-mediated SNARE disassembly [54], thus may actually limit or slow down the kinetics of membrane fusion. In fact, SNAP23, in the context of pancreatic islet β-cells, plays a similar role of an inhibitory SNARE for insulin granule fusion rather than a pro-fusion SNARE [35,55]. Perhaps a slower formation of autolysosomes is physiological (as in starvation) to allow downstream clearance (as a measure of maturation) of autolysosomes including any inadvertently-activated ZG proteases; and acceleration of this process by SNAP23 during pancreatitis stimuli might overwhelm these inefficient autolysosome clearance processes [13]. Whereas we showed that phosphorylation of SNAP23 enabled PM-bound SNAP23 to translocate to STX17-containing AVs, the actual mode of translocation could involve cytoskeleton proteins as was suggested in mast cells [45], but this will require further investigation. It would be most intriguing to see if novel SNAP23 blockers [35] could be used to treat AP, or alternatively, other strategies that could similarly slow down autolysosome formation.

Materials and Methods

Antibodies and reagents

Antibodies to STX3 (SySy, 110,033), STX4 (SySy, 110,042), SNAP23 (SySy, 111,202), VAMP2 (SySy, 104,202), VAMP8 (SySy, 104,302), LC3B (Cell Signaling Technology, 2775S), AMY/amylase (Santa Cruz Biotechnology, SC-46,657), ACTB/actin (Sigma, A1978), TUBA4A/tubulin (SySy, 302,211), FLAG (Sigma, F3165) and TAP (Antibodies Online, ABIN1173333), trypsin substrate Boc-Gln-Ala-Arg-AMC (Peptides International, MQR-3135-v) and commercial kits for AMY/amylase (Biovision, K711), lipase (Biovision, K722), MPO (Biovision, K744), CTSL (Abcam, ab65306) and CTSB (Abcam, ab65300) were described previously [15,44]. Anti-LAMP1and anti-SQSTM1/p62 antibodies were from Cell Science Technology (9091 and 39,749), and antibodies to SNAP29 and YKT6 were from Abcam (ab138500 and ab236583). Anti-STX17 antibody was from Sigma (HPA001204). Anti-Phospho-Ser-95 and Ser-120 SNAP23 antibodies that were fully characterized were from P. Roche [47], a coauthor. HRP conjugated anti-rabbit and anti-mouse secondary antibodies were from Jackson Immuno-Research (111–035-003 and 115–035-003). Anti-SNAP23 antibody for immunofluorescence [15] was from R&D Systems (AF6306). Fluorophore-conjugated secondary antibodies were from Jackson Immuno-Research (711–606-152) and R&D Systems (NL010).

Pancreas-specific SNAP23 depletion and pancreatitis induction

Ad-sc(scramble)-shRNA/RFP (WT) or Ad-SNAP23-shRNA/mCherry (SNAP23-KD) virus suspensions (300 µL), mixed with “India Ink” (to visualize distribution), were infused at a rate of 30 µL/minute into the pancreatic duct (Figure S6A,B) of male Sprague-Dawley rats (125–150 gm) to preferentially infect the pancreatic head and part of the body, sparing the tail, in a manner previously described, which effected optimal protein expression in a few days that lasted at least 2 months [48,49]. After 3 weeks, rats were subjected to AP induction by supramaximal caerulein (Sigma, C9026; intraperitoneal [i.p.] 20 μg/kg, 6-hourly injections) or a slightly modified acute ethanol pancreatitis protocol [56] (2-hourly i.p. injections of 1.35 gm/kg ethanol followed by 2-hourly 10 μg/kg i.p. caerulein injections). Control rats received equal volume of 0.9% saline or only-ethanol. Pancreases were harvested an hour after the last injection, and divided into head, body and tail, each further divided into 3 parts for functional, histological and biochemical studies described below.

Pancreas slicing, slice culture, ex vivo SNAP23 depletion and acini isolation, AR42J cell culture and transfection

Human pancreases (Table S1) were obtained from normal portions of pancreatic cancer operations and whole normal human pancreases from pre-terminal donors that were not used for transplantation and diverted to research (see Acknowledgments); both with approved IRBs. Human and rat pancreases were treated with 3.8% and 1.9% low-melting point agarose respectively in extracellular solution (ECS; in mM: 125 NaCl, 2.5 KCl, 1 MgCl₂, 2 CaCl₂, 1.25 NaH₂PO₄, 26 NaHCO₃, 2 sodium pyruvate, 0.25 ascorbic acid, 3 myo-inositol [Sigma, 15,125], 6 lactic acid [Sigma, L1750], 7 glucose) and sliced with a microtome to 100- to 120-µm-thick slices as described [15,44,57]. At least 10 slices per Petri dish were treated with SNAP23-KD (1.0 µL/mL for rat and 1.5 µL/mL for human) or WT (1 µL/mL) adenovirus for 2 h; 3–4 slices then placed on cell culture inserts (0.4 μm, 30 mm diameter, EMD Millipore) coated with 3 mg/ml type-1 rat tail collagen. Pancreatic slices were used for amylase secretion, immunoprecipitation and western blot assays. For imaging, 48-hr cultured slices from the culture inserts were dispersed into acini by a widely- used mechanical and enzymatic dissociation technique [15,44] and plated on poly-L-lysine-coated coverslips. A rat pancreatic acinar cell line (AR42J) was maintained in RPMI1640 media (with 10% fetal bovine serum, 100 units/ml penicillin and streptomycin, 2 mM L-glutamine) in humidified cell culture incubator at 37°C with 5% CO₂ [58]. For transfection, cells were seeded at 70% confluency in 60-mm tissue culture dishes (Corning, 430,166). After 24 h, cells were transiently transfected with 4 µg of N-terminal FLAG-tagged (in pCMV-Tag2B; Stratagene, 211,172–52) WT-SNAP23 or IKBKB/IKKβ phosphorylation-disabled mutant of SNAP23 (S95A S120A) using AR42J transfection reagent (Altogen Biosystems, 1180). Twenty-four h post-transfected cells were replaced with fresh media with dexamethasone (100 nM; Cell Signaling Technology, 14,776) to induce acinar differentiation over 48 h [59], washed (2X) with ECS and then stimulated with 10 nM CCK or 100 pM CCK plus 20 mM EtOH in 5 mL ECS for 30 min.

Enzyme assays

For amylase secretion assay, 3–4 pancreatic slices in Krebs-Ringer-HEPES buffer (KRH; in mM: NaCl 104, KCl 5, KH₂PO₄ 1, MgCl₂ 1.2, HEPES 25, CaCl₂ 1.2, D-glucose 2.5, L-glutamine 2; supplemented with MEM [Invitrogen, 11,130–051 and 11,140–050], 0.015% [w:v] soybean trypsin inhibitor [Worthington, S9B11099], and 0.2% [w:v] BSA [Sigma, A7030]; pH 7.4) were stimulated with CCK at 37°C for 1 h; secreted amylase then determined by colorimetric assay [15,44]. Serum amylase, lipase, pancreatic MPO, lysosomal CTSL and CTSB activities were determined using commercially-available kits mentioned in “Antibodies and reagents” section [15,44]. Trypsin activities in pancreatic tissue homogenates and acinar lysates were measured by the release of 7-amino-4-methylcoumarin (AMC) using a fluorimeter (Fluostar Optima, Isogen Life Science, De Meern, Netherlands) [15,44].

Immunoprecipitation (IP) and western blot

Two hundred µg debris-free (18,000 g for 10 min, 4°C) pancreatic slice or AR42J homogenates (in lysis buffer, pH 7.4: 25 mM HEPES, 100 mM KCl, 1.5% Triton X-100 [Sigma, T8532]), protease inhibitors (Cell Signaling Technology, 5872S) were precleared with 50 μL of protein-A sepharose (Invitrogen, 101,041) followed by IP with 2 µg antibody conjugated protein-A sepharose as we had described [15,44]. IPed proteins were extracted by boiling with 20 µL SDS-PAGE sample buffer (2 M Tris-HCl, pH 6.8, 20% SDS, 30% glycerol, 0.03% phenol red). Secreted amylase in buffer (500 µL) were harvested by overnight acetone (4X, v:v) precipitation at −20°C followed by centrifugation for 5 min at 18,000 g at 4°C and boiling in 30 µL sample buffer. IP-eluted proteins, acetone precipitated proteins and tissue lysates, all in sample buffer were separated on 12–15% gradient SDS-PAGE and identified by specific antibodies. Western blot intensities of the identified proteins, including loading controls were quantified within the linear range using ImageJ 1.49 t software from http://rsbweb.nih.gov/ij/download.

Spinning-disk confocal microscopy

All images were taken at a magnification of 94.3X (63X oil immersion objective, 1.35 NA) with a Zeiss spinning disk confocal microscope equipped with (in nm) 405, 491 561, and 605 lasers (Spectral Applied Research, Concord, ON, Canada), and a Hamamatsu C9100-13 EM-CCD camera (Hamamatsu Photonics, Shizuoka, Japan). Volocity 3DM software (Perkin Elmer Corporation, Waltham, MA) was used to drive the system and to analyze images [15,44].

Statistical analysis

Statistical analyses were performed using 2-tailed Student’s t-test or one-way analysis of variance in ORIGIN (Microcal, Amherst, MA). Data are presented as means±SEM. P value <0.05 was considered significant.

Funding Statement

This work was supported by a grant from Canadian Institute of Health Research (CIHR PJT-159542) to H.Y.G. and a “Research Grant Fellowship” from the American Pancreatic Association/National Pancreas Foundation to S.D. Some of the equipment used in this study was supported by the 3D (Diet, Digestive Tract and Disease) Centre funded by the Canadian Foundation for Innovation and Ontario Research Fund, project number 19442.

Disclosure statement

All authors declare no competing financial interests.

Ethics approval

Animal procedures and use of human pancreas were performed in accordance with the University of Toronto’s Animal Care Committee’s ethical guidelines and Research Ethics Board of the University Health Network of Toronto, and with approved IRBs.

Supplementary material

Supplemental data for this article can be accessed here.