Main

Cystic fibrosis (CF) is a genetic disease caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene, which encodes a cAMP-regulated chloride (Cl-) channel(1,2). As a consequence of absent or defective cAMP-regulated Cl- conductance, various exocrine organs lined by epithelial tissue become clogged with thick, inspissated secretions. In the CF pancreas, obstruction of small pancreatic ducts by proteinaceous secretions has been implicated in the pathogenesis of pancreatic disease(3,4). Reduced fluid and ductal chloride and bicarbonate secretion have been observed in both pancreatic-sufficient and pancreatic-insufficient patients with CF(5–7). Defective anion secretion has been attributed to absent CFTR, or its dysfunction, within pancreatic ductal cells(5,7).

Studies of CF infants show high circulating concentrations of pancreatic enzymes (trypsinogen and lipase), irrespective of pancreatic function status(8). In patients with CF and pancreatic insufficiency, serum enzyme concentrations drop, reaching immeasurable or low values by 5-7 y of age(9,10). This observation almost certainly reflects progressive destruction and atrophy of acinar tissue. In contrast, CF patients with pancreatic sufficiency have widely fluctuating serum trypsinogen concentrations at all ages, ranging from normal to values greatly exceeding the normal range(9,10). High serum concentrations of pancreatic enzymes could be the result of obstruction of small pancreatic ducts leading to regurgitation of enzymes into the circulation. Alternatively, high circulating concentrations of pancreatic enzymes may be the result of a CF-mediated defect in exocytosis within acini.

Mouse models of CF have been developed by targeted disruption of CFTR gene locus in embryonic stem cells(11,12). Mouse models (Cftr-/-) show severe gastrointestinal complications similar to those seen in CF patients(11). Unlike humans, these Cftr-/- mice exhibit only minor focal histologic abnormalities in the exocrine pancreas(11,12). This is attributable to the existence of a Ca2+-regulated chloride conductance in pancreatic ductal tissue of Cftr-/- mice(13,14). However, recent studies of pancreatic acini from Cftr-/- mice provide evidence for involvement of the acinar cells in development of CF pancreatic pathology. De Lisle(15) demonstrated that the pancreatic acinar tissue was affected, as evidenced by dilated acinar lumina. The glycoprotein gp300, which is mainly localized to the zymogen granule membrane in normal tissue, was found to line the distended luminal membranes in CF tissue. The increased expression of gp300 in Cftr-/- mice may cause poorly soluble exocrine protein secretion, contributing to the development of pathology in the mouse pancreas(15). Freedman et al.(16–18) demonstrated that acidic medium (pH < 6.0) largely abolished apical endocytosis in rat pancreatic acini, and the pancreatic acini mimicked the morphologic findings observed in acinar cells of CF patients. Studies of Cftr-/- mice showed markedly impaired endocytosis in pancreatic lobules, as assessed by horseradish peroxidase uptake(19). These observations indicate that pathophysiologic changes within acini may contribute to the pathogenesis of CF pancreatic dysfunction. Therefore, it is an ideal model to study pancreatic acini.

A dietary method developed by us to prolong the survival of Cftr-/- mice with a severe intestinal phenotype(20) afforded the opportunity to evaluate the amylase secretory response of isolated pancreatic acini using a variety of secretagogues involved in exocytosis. We recently demonstrated that pancreatic enzyme activities of Cftr-/- mice are significantly lower than age-matched controls fed a solid or liquid diet(21). However, it remains unclear whether enzyme secretory response to secretagogues is altered in the Cftr-/- mouse. Evidence of cAMP-regulated Cl- conductance within rabbit acini(22,23), identification of CFTR proteins in the luminal membrane of rat and mouse acinar cells(24), and strong dependence of enzyme secretion on chloride conductance in acinar cells(25,26) led us to hypothesize that changes in Cl- permeability within acini may be associated with alterations in enzyme secretion in the CF pancreas.

MATERIALS AND METHODS

Chemicals and reagents. 8-Br-cAMP was purchased from Boehringer Mannheim GmbH (West Germany). Forskolin, Soybean trypsin inhibitor (SBTI), and carbachol were obtained from Sigma. Purified collagenase was from Worthington Biochemicals Corp. (Freehold, NJ). Phadebas amylase reagent tablets were from Kabi Pharmacia Diagnostics AB (Uppsala, Sweden), and Eagle medium amino acid supplement was from GIBCO Life Technologies, Inc. All other chemicals and reagents used in this study were of analytical grade.

Breeding and care of Cftr-/- mice. Cftr-/- mice were bred and cared for according to a previously described protocol(11,20). Mice identified as controls (Cftr+/+) were weaned about 21 d to solid chow or to a liquid diet (Liquidet, Bio-Serv, Frenchtown, NJ). Cftr-/- mice were weaned to the liquid diet exclusively. The dietary composition of the solid and liquid diets, respectively, was protein 14 and 14%, fat 6 and 5%, fiber 4.5 and 1.5%, ash 8 and 5%, carbohydrate 63 and 69%. The protein in the liquid diet was hydrolyzed casein.

Preparation of acini. Animals were fasted overnight before sacrifice. Pancreatic acini were prepared according to Williams et al.(27) with modification(28). Six mice, aged 5-6 wk, were used in each experiment. Following sacrifice by cervical dislocation, each pancreas was quickly removed and placed in a standard solution (pH 7.4) containing 98 mM NaCl, 6 mM KCl, 2 mM KH2PO4, 5 mM pyruvate, 6 mM fumarate, 5 mM glutamate, 11.5 mM glucose, 25 mM HEPES, 2 mM CaCl2, 1.2 mM MgCl2, 0.01% soybean trypsin inhibitor (SBTI), 2% (vol/vol) essential amino acid mixture (×100), 1% essential vitamin mixture (×100), 2 mM glutamine, 0.2% BSA. Pancreata were trimmed of fat and connective tissue, injected with a total of 5 mL of digestion solution [standard solution with 45 U/mL purified collagenase (∼1000 U/mg)], placed in a plastic flask at room temperature, gassed with 100% O2, and incubated at 37°C in an oscillating water bath for 20 min. The digestion solution was then decanted and replaced with fresh buffer and enzyme. The pancreata were incubated at 37°C for an additional 20 min. Following digestion, the acini were sequentially pipetted through 10-mL and 5-mL plastic pipets. Acini were then purified by filtration through a 150 µM steel sieve and centrifugation in 4% BSA (three times). Acini were resuspended in the standard incubation medium with 0.5 mM CaCl2 and 1% BSA and gassed with 100% O2.

Secretagogue-stimulated amylase secretion. Preliminary experiments were conducted to evaluate maximal effects on amylase secretion. On the basis of these experiments, concentrations of secretagogues used in this study were 1 mM 8-Br-cAMP, 0.2 mM forskolin, and 1 µM carbachol. The maximal concentration of carbachol used in this study was consistent with previous observations of Williams et al.(27) and De Lisle et al.(29). Carbachol was dissolved in distilled water while all other reagents were dissolved in DMSO on the day of experiment. The final DMSO concentration never exceeded 2%, which had no significant effects on basal amylase release in the preliminary experiments.

Following preincubation at 37°C for 60 min., the acini were pelleted by low-speed centrifugation (150 × g for 1 min) and then resuspended in the fresh incubation medium. Aliquots (0.5 mL) of acinar suspension (0.5 mg/mL) were placed in 1.5-mL microfuge tubes. Amylase secretory responses were measured at 45 min (37°C) following addition of secretagogues. Following incubation, the aliquots were centrifuged at 11,000 rpm for 30 s using a Beckman Microfuge. The supernatant was added to a diluting solution (pH 7.8) containing 10 mM NaH2PO4, 0.1% SDS, and 0.1% BSA and saved for determination of amylase release. Basal amylase release was determined at 45 min in the absence of secretagogues. Total cellular amylase activity was determined following sonication of 50-µl aliquots of the acinar suspension diluted with 2.45 mL of the diluting solution. Secretagogue-stimulated amylase secretion was calculated by subtracting basal amylase release from the stimulated amylase concentration after secretagogue stimulation.

Simultaneous experiments were performed with Ca2+- and cAMP-mediated secretagogues alone, or in combination, using the same preparation of pancreatic acini. We then compared the amylase secretion following secretagogues given in combination with the calculated sum of values produced by each agonist alone. This approach, therefore, permits direct comparisons of the combined and individual secretagogue effects on amylase secretion under identical conditions, which would serve to eliminate ambiguities produced by separate measurements.

Amylase activity, protein, and DNA content. Amylase activity was measured by the method of Ceska et al.(30) using Phadebas reagent tablets. Protein was determined by the method of Lowry et al.(31). DNA was measured by fluorometric quantification as described by Downs and Wilfinger(32), using calf thymus DNA as standard.

Data analysis. Data are reported as means ± SD of six to eight experiments in each group. Each experiment was performed in triplicate using an independent acinar preparation. Differences between the means of two groups were determined using the unpaired two-tailed t test unless otherwise stated. Amylase activities are expressed both as a percentage of intracellular acinar amylase content and as unit of enzyme activity per mg DNA. The results were considered to be significantly different at p < 0.05.

RESULTS

Effects of diet and CF on somatic and pancreatic growth. As shown in Table 1, the balanced liquid diet had no effects on somatic and pancreatic weight of the controls. However, consistent with previous observations(20), both body weight and pancreatic weight were significantly reduced in Cftr-/- animals in comparison with the age-matched Cftr+/+ controls fed either a solid or a liquid diet (p < 0.001 and <0.05, respectively). There were no differences in pancreatic:somatic weight ratios between the three groups.

Table 1 Comparison of body weight, pancreatic weight, and pancreatic to body weight ratio in Cftr+/+ and Cftr-/- mice
Full size table

Alterations in pancreatic DNA content and acinar amylase activity. Pancreatic DNA and acinar amylase content of the two Cftr+/+ control groups and Cftr-/- animals are shown in Table 2. There were no significant differences in pancreatic DNA content between Cftr-/- mice and Cftr+/+ control animals fed by solid or liquid diet. Liquid-fed Cftr+/+ mice showed lower acinar amylase content than the age-matched solid-fed Cftr+/+ controls (p < 0.05). Acinar amylase activity of the Cftr-/- animals was significantly lower than the solid-fed (p < 0.005) and the liquid-fed Cftr+/+ animals (p < 0.05). When expressed as U/mg DNA, the acinar amylase content of Cftr-/- mice was not different to the liquid-fed Cftr+/+ controls but was significantly lower than the solid-fed Cftr+/+ animals (p < 0.05).

Table 2 Pancreatic DNA and acinar amylase activity in pancreatic acini of the Cftr+/+ and Cftr-/- mice
Full size table

Secretory response of acini from Cftr+/+ mice. We first evaluated amylase release following stimulation by cAMP- and Ca2+-mediating secretagogues (individually and in combination) in acini from conventionally fed Cftr+/+ mice. When expressed as a percentage of acinar amylase content, 1 mM 8-Br-cAMP and 0.2 mM forskolin stimulation exhibited a similar secretory response (5.5 ± 3.9% and 5.8 ± 3.5%, respectively), but each secretagogue induced a significantly higher response than 1 µM carbachol alone (1.7 ± 0.8%, n = 7, p < 0.05). BrcAMP and carbachol, given in combination, produced an additive effect on enzyme secretion, as the sum of amylase secretion induced by BrcAMP and carbachol individually was similar to that observed when both secretagogues were used in combination (7.2% versus 6.9%, not significant). The combination of BrcAMP and forskolin also produced an additive effect on amylase secretion in the Cftr+/+ acini (9.7% versus 9.1%, n = 6).

In the Cftr+/+ animals fed a liquid diet (Fig. 1), net amylase secretion in response to the various secretagogues was no different to that observed in the conventionally fed animals (Fig. 1A). Amylase secretion, expressed per mg of DNA, tended to be lower in the liquid-fed Cftr+/+ controls, but the results were not significantly different (Fig. 1B). To obviate any dietary effects, the acinar secretory response data from the liquid-fed Cftr+/+ animals were compared with those obtained from the age-matched Cftr-/- mice.

Figure 1
figure 1

Effects of diet on amylase secretion. Data were expressed as a percentage of acinar amylase content (A) and net amylase secretion/mg DNA (B). The liquid diet had no effect on amylase secretion in response to the various secretagogues (given alone or in combination) when expressed as a percentage of acinar amylase content (p > 0.05). Net amylase secretion (U/mg DNA) tended to be lower in the liquid-fed Cftr+/+ controls, but the differences were not significant (p > 0.05).

Full size image

Secretory response of acini from Cftr-/- mice. When expressed as a percentage of acinar amylase content or per mg of DNA, there were no differences in amylase secretory response to the various secretagogues (given alone or in combination) between the liquid-fed Cftr+/+ and Cftr-/- mice (Fig. 2). However, there was a tendency for an enhanced acinar secretory response (% of total amylase content) in Cftr-/- mice following all treatments except BrcAMP alone (Fig. 2A).

Figure 2
figure 2

Amylase secretion in the liquid-fed Cftr+/+ and Cftr-/- mice. When expressed as a percentage of acinar amylase content (A), no differences were observed in amylase secretory responses between the liquid-fed Cftr+/+ and Cftr-/- mice. Similarly, no significant differences in secretagogue-induced amylase secretion expressed per mg of DNA (B) were observed between the liquid-fed Cftr+/+ and Cftr-/- mice.

Full size image

As observed in the liquid-fed Cftr+/+ mice, the combination of forskolin and BrcAMP induced an additive effect on amylase secretion in acini from Cftr-/- mice. That is, the calculated sum of amylase secretion following BrcAMP and forskolin each given alone (12.4% or 323.7 U/mg DNA) was not significantly different to that observed when given in combination (12.8% or 334.1 U/mg DNA).

We compared amylase secretion following BrcAMP and carbachol with the calculated sum of values produced by BrcAMP and carbachol given alone in the Cftr+/+ and Cftr-/- groups. In the liquid- and solid-fed Cftr+/+ animals, no significant differences in the secretory response were observed between the combined effects of these agonists and the calculated response induced by BrcAMP and carbachol given individually (Table 3). In contrast, BrcAMP given with carbachol produced a significantly greater amylase secretory response in Cftr-/- mice, which was 32% greater than the calculated sum of the secretory response of each agonist given alone (p < 0.05). Thus, BrcAMP and carbachol induced a synergistic effect on amylase secretion in pancreatic acini from Cftr-/- mice but not in the Cftr+/+ animals.

Table 3 Effect of BrcAMP and carbachol on amylase secretion (% of acinar amylase content) in Cftr+/+ and Cftr-/- mice
Full size table

DISCUSSION

Exocrine secretion, which involves the insertion of cytoplasmic secretory granules into the plasma membrane, requires the coordinated action of both membrane fusion and electrolyte transport. Fluid secretion by acinar cells relies on the presence of chloride channels in the luminal membrane(33,34). In addition, the pancreatic zymogen granule membrane contains chloride and cation channels, such as an ATP-sensitive K+ conductance pathway, which promote local fluid production to solubilize and flush the macromolecular contents of the granule into the acinar lumen(33,34). Kopelman and Gauthier(22) identified a cAMP-activated chloride conductance within rabbit pancreatic acini and, in a subsequent study, demonstrated expression of CFTR as a cAMP-activated chloride pathway(23). Recently, Zeng et al.(24) provided direct evidence for expression of CFTR protein and CFTR-dependent chloride channels in the luminal membrane of rat and mouse pancreatic acinar cells. Furthermore, previous studies have shown that enzyme secretion from permeabilized acini of the rat pancreas evoked by cholecystokinin, carbachol, and second messengers (phorbol ester plus cAMP) is strongly dependent on the presence of Cl- in the surrounding medium(25). Secretion was reduced by Cl- and K+ channel blockers and by hyperosmotic solutions(25). Pretreatment of animals with the secretory agonists secretin or cholecystokinin resulted in activation of Cl- transport pathway in the membrane of rat pancreatic zymogen granules(26).

Taken together, these observations suggest that chloride transport pathways may play a central role in stimulus-secretion coupling of enzymes in pancreatic acini(25,26). We hypothesized defective or absent CFTR expression in acini may influence the enzyme secretory response to secretagogues. Thus, Cftr-/- mice, lacking expression of the CFTR protein(11,35), would be expected to show a reduced secretory response to cAMP or cAMP-mediated agonists. However, our studies of cAMP-mediated amylase release of isolated acini showed an identical secretory response in Cftr-/- and Cftr+/+ animals (Fig. 2A), which suggests that CFTR-mediated mechanisms are not essential for pancreatic enzyme secretion. Although amylase secretion expressed per mg of DNA was lower in acini of the Cftr-/- mice (Fig. 2B), total amylase content (U/mg DNA) was also reduced in comparison with the diet-matched Cftr+/+ controls. Similarly, carbachol-induced amylase secretion, a calcium-mediated mechanism of enzyme secretion, was not altered in Cftr-/- acini. Consistent with these findings, Kopelman et al.(23) demonstrated that rabbit acini, incubated with CFTR antisense, demonstrated a loss of CFTR mRNA, CFTR protein, and BrcAMP-activated chloride efflux without affecting BrcAMP-induced amylase release. Two recently published abstracts also showed no significant differences in Ca2+- or BrcAMP-induced amylase secretion between controls and CF mice(19,36).

Alternatively, the normal cAMP-mediated amylase secretion in Cftr-/- acini may be mediated by Ca2+-activated chloride channels in the luminal membrane(34,37). A rise in free cytosolic Ca2+ concentration in the region of the secretory granule not only activates exocytosis but also induces gating of Ca2+-sensitive chloride channels in the luminal membrane inducing acinar fluid secretion(37,38). Studies of permeabilized mouse pancreatic acini support the role of Ca2+ as an intracellular mediator of enzyme secretion. In the absence of Ca2+, cAMP has no effect on amylase secretion(39), whereas the presence of Ca2+ in the incubation medium significantly enhances cAMP-induced amylase secretion. Furthermore, in CF mouse tracheal epithelia, forskolin increases intracellular Ca2+-mediated chloride secretion(40). Thus, it would be reasonable to speculate that Ca2+-sensitive chloride channels in the luminal membrane play an alternative role in enzyme secretion and, at least partly, account for the normal cAMP-induced amylase secretion in Cftr-/- acini.

Interestingly, control acini showed a significantly higher amylase response to forskolin and BrcAMP than carbachol (p < 0.05). This observation suggests that cAMP-mediated pathways play a central role in enzyme secretion, which may be a unique feature for this strain of mice. Net maximal carbachol-induced amylase release was less than 2%, which is lower than that observed in previous reports. Net amylase release of 5% was observed in one study(41), whereas 7% released was seen with permeabilized mouse acini(29). Both studies used White Swiss mice. We cannot fully explain for lower amylase secretion in our experiments, but experimental conditions and varying mouse strains may account for the differences. Furthermore, the reduced amylase content in Cftr-/- mice is consistent with previous observations by De Lisle(15) and our own laboratory(21). We have previously suggested the reduced amylase content in acini from Cftr-/- mice may result from a primary effect of loss of CFTR function, but in addition, decreased amylase synthesis owing to malnutrition may be responsible(42).

The combination of BrcAMP and carbachol induced a synergistic effect on net amylase secretion in Cftr-/- acini; synergism was not observed in acini from either solid-fed or liquid-fed Cftr+/+ mice. Although the mechanisms remain unclear, loss of CFTR function may induce an amplified secretory response via an alternative mechanism. Alternatively, this observation may be a secondary effect of malnutrition because the Cftr-/- mice were significantly underweighted in comparison with the diet-matched controls and had reduced pancreatic weight and acinar amylase content. In a rodent model of malnutrition, Schick et al.(43) observed acinar atrophy and a decrease in number and size of secretory granules, which was strikingly similar to that seen in acinar cells of Cftr-/- mice(15).

Regardless of the cause, the enhanced secretory response in CF acini may be of pathophysiologic significance and may be implicated in the development of pancreatic dysfunction in CF patients. First, in the presence of defective fluid secretion(24), enhanced secretion of pancreatic enzymes may contribute to the development of pancreatic pathology in CF patients by facilitating intraluminal precipitation and occlusion of hyper-concentrated proteinaceous secretions within acinar lumina and small pancreatic ducts. Second, the amplified secretory response would explain the increased serum pancreatic enzyme concentrations observed in young CF patients, which is seen irrespective of pancreatic function status(8). CF patients with pancreatic sufficiency also have high, widely fluctuating serum trypsinogen concentrations at all ages(9,10). An in vivo study of anesthetized rats demonstrated that circulating pancreatic enzyme concentration secreted via the constitutive pathway can be increased by cholecystokinin (CCK-8) stimulation(44). Thus, enhanced pancreatic enzyme secretion via constitutive pathway (across basolateral membranes) may contribute to the increased circulating concentration of pancreatic enzymes.

Regardless of the mechanism, the enhanced secretory response in Cftr-/- mice almost certainly involves an amplified intracellular signal transduction process. Several direct and indirect observations from our study support this contention. 1) The additive effects of BrcAMP and carbachol on amylase secretion in acini from the wild-type animals suggest that at least two intracellular pathways act in parallel(39). 2) In acini from Cftr-/- mice, the amylase response to each agonist given alone was unaltered, which suggests that cAMP- and Ca2+-mediated signal transduction pathways remain intact. 3) The synergistic effect of amylase secretion by BrcAMP and carbachol, which was only observed in acini from Cftr-/- mice, suggests that intracellular signal transduction pathways are amplified at a step distal to cAMP- and Ca2+-mediated events.

Freedman et al.(16–18) have proposed that pancreatic dysfunction in CF involves the defective coupling of both ductal and acinar cell function in the exocrine pancreas. Defective chloride and chloride-bicarbonate exchange induces progressive acidification of the acinar and ductal lumen, leading to secondary defects in apical trafficking (endocytosis) of zymogen granule membranes and solubilization of secretory proteins. This hypothesis is based on the observation that low luminal pH affects endocytosis and release of glycoprotein-2, a major glycoprotein within the zymogen granule membrane(16,17). Provided the buffer medium is between pH 7.4 and 8.3, cholecystokinin stimulation of isolated rat pancreatic acini induces luminal membrane contraction and cleavage of glycoprotein-2 from the apical cell surface. In contrast, under acidic conditions (pH < 6.25), there is poor release of glycoprotein-2, and the acinar lumen becomes dilated and filled with aggregated proteins, suggesting impairment in membrane internalization and solubilization of secretory proteins(16). In a subsequent study, these investigators demonstrated reduced rate of endocytosis in the pancreatic lobules of Cftr-/- mice(19), which may be because of low intraluminal pH owing to defective chloride secretion and reduced bicarbonate-chloride exchange. Morphologic studies by De Lisle(15) also demonstrated that acinar lumina of the pancreatic lobules from Cftr-/- mice are greatly dilated and filled with aggregated protein. Increased expression of gp300 in Cftr-/- mice may cause poor solubilization of the secretory proteins. None of these observations preclude the possibility that enhanced exocytosis from acinar cells of Cftr-/- mice may contribute to disease pathogenesis.

In conclusion, multiple derangements within ductal and acinar cells may contribute to the pathogenesis of pancreatic disease in CF. Defective chloride and chloride-bicarbonate exchange within epithelial cells in small ducts and centro-acinar cells will reduce ductal fluid secretion and impair alkalinization of the acinar lumen. Acidic intraluminal pH induces defective recycling of secretory membranes derived from zymogen granules, leading to massive dilatation of the acinar lumen, loss of the apical pole of the acinar cell, and a marked reduction in the number of zymogen granules(18). Finally, our observation of an enhanced secretory response to a combination of agonists, in the murine pancreas, may be of clinical relevance to the pathogenesis of pancreatic disease in CF, as increased secretion of exportable proteins would increase the risk of microprecipitation in acinar lumina and small ducts.