Cell Physiology

A role for Rho and Rac in secretagogue-induced amylase release by pancreatic acini

Yan Bi, John A. Williams


The actin cytoskeleton has long been implicated in protein secretion. We investigated whether Rho and Rac, known regulators of the cytoskeleton, are involved in amylase secretion by mouse pancreatic acini. Secretagogues, including cholecystokinin (CCK) and the acetylcholine analog carbachol, increased the amount of GTP-bound RhoA and Rac1 and induced translocation from cytosol to a membrane fraction. Immunocytochemistry revealed the translocation of Rho and Rac within the apical region of the cell. Expression by means of adenoviral vectors of dominant-negative Rho (RhoN19), dominant-negative Rac (RacN17), and Clostridium Botulinum C3 exotoxin, which ADP ribosylates and inactivates Rho, significantly inhibited amylase secretion by CCK and carbachol; inhibiting both Rho and Rac resulted in a greater reduction. This inhibitory effect of RhoN19 on CCK-induced amylase secretion was apparent in both the early and late phases of secretion, whereas RacN17 was more potent on the late phase of secretion. None of these three affected the basal Ca2+ or the peak intracellular Ca2+ concentration stimulated by CCK. Latrunculin, a marine toxin that sequesters actin monomers, time-dependently decreased the total amount of filamentous actin (F-actin) and dose-dependently decreased secretion by secretagogues without affecting Ca2+ signaling. These data suggest that Rho and Rac are both involved in CCK-induced amylase release in pancreatic acinar cell possibly through an effect on the actin cytoskeleton.

  • cholecystokinin
  • carbachol
  • pancreas
  • cytoskeleton

at least 20 Rho family member proteins are known; RhoA, Rac1, and Cdc42 are the best characterized. In fibroblasts, RhoA, Rac1, and Cdc42 regulate the actin cytoskeleton by stimulating the formation of stress fibers and focal adhesions, lamellipodia and filopodia, respectively (31, 39, 40). Like other members of the Ras small GTPase family, Rho proteins function as molecular switches to transduce extracellular stimuli to intracellular responses and cycle between a soluble, GDP-bound inactive state and a membrane-associated GTP-bound active state. At rest, most Rho proteins are mainly located in the cytoplasm as a complex with RhoGDI, and after stimulation, activated Rho translocates to membranes and there interacts with downstream effectors to regulate several cell functions (1, 44). To date, >40 effectors have been described for the mammalian Rho family, including Rho kinase, protein kinase N, several phosphatidylinositol kinases, PLC, phospholipase D, and mDia (3, 8, 29, 48). Among these effectors, Rho kinase is the best understood as a regulator of smooth muscle contraction. In addition to actin reorganization, Rho family members have been also associated with gene transcription, the regulation of cell polarity, G1 cell cycle progression, microtubule dynamics, a variety of enzymatic activities, and vesicular transport pathways (5, 12, 16, 27).

Rho, Rac, and Cdc42 have been implicated in the regulation of exocytosis in various cells, including pancreatic beta cells (21), basophilic leukemia cells (36), mast cells (6), parietal cells (45), PC-12 cells, and neurons (11). Rac has been shown to be associated with synaptic vesicles in neurons (11) and RhoA is localized on secretory granules in chromaffin cells (14, 15). Although increasing evidence implicates the involvement of Rho family proteins, the exact functions of Rho proteins in secretion are still unclear.

Pancreatic acinar cells synthesize and secrete digestive enzymes in response to cholecystokinin (CCK) and ACh and have long served as model cells to study secretion. Zymogen granules (ZGs) containing digestive enzymes such as amylase are stored in the apical region of the acini. The binding of CCK and ACh to their receptors on the basolateral membrane of the acini activates heterotrimeric G proteins of the Gq family, which then activate a PLC-β, which in turn cleaves the membrane phospholipid PIP2 to inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (50, 53). IP3 then binds to its receptors on the endoplasmic reticulum and increases calcium release, which in turn triggers exocytosis to release the enzyme contained in ZGs. Before ZGs can fuse with the apical membrane, they have to cross a band of actin filaments underneath the apical membrane, termed the “terminal web.” Under resting conditions, the actin cytoskeleton prevents secretory granules from reaching their exocytic destination and can act as a physical barrier. Phalloidin, which polymerizes actin and prevents its disassembly, blocks stimulated secretion in permeabilized pancreatic acinar cells (28). In addition, disruption of the actin cytoskeleton with cytochalasin or actin monomer binding proteins inhibits stimulated secretion (7, 28, 49), suggesting that the actin cytoskeleton can also play a positive role in secretion. Visible disassembly of apical actin during supramaximal secretagogue stimulation has been reported and is accompanied by inhibition of secretion (7, 19, 34). Local actin reorganization and actin coating of ZGs have also been observed under physiological conditions (30, 47). There is limited information available regarding the activation and function of Rho and Rac in pancreatic acinar cells, although previous studies (22) of NIH3T3 cells stably transfected with CCKA receptor have demonstrated that Rho can be activated by stimulating the CCKA receptor. Recently, Rho has also been implicated in CCK-induced acinar morphological changes, which could be blocked with C3 toxin in permeabilized acinar cells (20). The aim of this study, therefore, was to determine whether Rho and Rac can be activated by CCK and whether they are involved in the amylase release evoked by secretagogues.



Collagenase was purchased from Worthington (Lakewood, NJ) and Crescent Chemical (Islandia, NY); sulfated CCK octapeptide was from Research Plus (Bayonne, NJ), and the Rho-kinase inhibitor Y-27632, was purchased from BioMol (Plymouth Meeting, PA). Triple hemaglutinin antigen (HA)-tagged dominant negative Rho (RhoN19) and dominant negative Rac (RacN17) in the PKH3 plasmid (26) were obtained from Ian Macara (University of Virginia, Charlottesville, VA) and C3 plasmid was from Silvio Gutkind (NIH, Bethesda, MD). Glutathione-S-transferase (GST)-Rhotekin RhoA binding domain (RBD) and GST-p21-activated kinase (PAK) p21-binding domain (PBD) protein were purchased from Cytoskeleton (Denver, CO) and DMEM from Invitrogen (Carlsbad, CA). Mouse monoclonal and polyclonal anti-RhoA antibodies were purchased from Santa Cruz (Santa Cruz, CA), monoclonal anti-Rac1 was from Upstate (Chicago, IL), and monoclonal anti-HA (12CA5) was from Roche Molecular Biochemicals (Indianapolis, IN). Femto Western blotting reagent was purchased from Pierce (Rockford, IL). All other chemical reagents were obtained from Sigma.

Construction of recombinant adenoviruses.

The recombinant adenoviruses encoding RhoAN19, Rac1N17, and C3 exotoxin were constructed according to the method of He et al. (17). Briefly, cDNAs encoding the HA-tagged small G proteins and C3 were cloned into pAdTrack-cytomegalovirus (CMV). The resultant plasmid was then transformed into AdEasier-1 cells (BJ5183 derivatives, which already contain the AdEasy-1 plasmid). Recombinants were selected and confirmed by restriction endonuclease assay and DNA sequencing. Finally, linearized recombinant plasmid was transfected into human embryonic kidney-293 cells for packaging. Recombinant adenoviruses were collected 7–12 days after infection and purified by CsCl gradient ultracentrifugation. The titers of the viral stocks were estimated by counting EGFP-expressing cells as pAdTrack-CMV also encodes EGFP driven by a separate CMV promoter. An adenovirus expressing β-galactosidase and EGFP was used as a control. The adenovirus expressing Cdc42N17 was purchased from the University of Iowa Gene Transfer Vector Core (Iowa City, IA).

Preparation, short-term culture, and viral infection of pancreatic acini.

Acini were prepared by methods previously described (52, 54). In brief, pancreata were excised from freely fed adult male ICR mice weighing 22–27 g. Acini were isolated by enzymatic digestion with collagenase, followed by mechanical shearing. The acini were then filtered through 150-μm Nitex mesh, purified by sedimentation through 4% BSA in buffer, and suspended in a HEPES-Ringer buffer (HRB) containing 5 mg/ml BSA and 0.1 mg/ml soybean trypsin inhibitor. For immediate study, acini were kept in HRB at 37°C while being shaken at 35 cycles/min. For overnight incubation, cells were cultured without shaking at low density in 10 cm petri dishes in DMEM enriched with 0.5% bovine serum albumin, 0.01% soybean trypsin inhibitor, and antibiotics, and incubated at 37°C with 5% CO2 for 16 h. For the viral infection experiments, 107 pfu/ml adenoviruses were added to the culture medium at the beginning of the incubation. The morphological state of acini, either freshly prepared or after overnight culture, was assessed by microscopic examination of 0.5-μm-thick sections prepared after being embedded in resin and stained with toludine blue (49).

Analysis of agonist-stimulated amylase secretion.

After incubation for 16 h with adenovirus or preincubation for 30 min for fresh acini, acini were allowed to settle by gravity, resuspended in HRB, and then incubated with various concentrations of CCK or carbachol at 37°C in 1-ml aliquots in plastic vials that were shaken at 50 cycles/min. After the specified time, the acinar suspension was centrifuged for 30 s in a microcentrifuge and the supernatant assayed for amylase activity with the use of Infinity Amylase reagent (Thermo Electron, Woburn, MA) in a plate reader (μQuant, Bio-Tek Instruments, Winooski, VT). Results were expressed as a percentage of initial acinar amylase content.

Rho and Rac pull-down assay.

Active Rho and Rac were determined by a pull-down assay according to a modified method from Ren et al. (38). Briefly, pancreatic acini were stimulated as noted and then collected in 300 μl of lysis buffer (50 mM Tris, pH 7.5, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 500 mM NaCl, and 10 mM MgCl2) and centrifuged to remove cellular debris. The supernatant was then mixed with GST Rhotekin RBD beads or GST PAK-PBD beads and rotated at 4°C for 45 min. The beads were then centrifuged, washed multiple times, and the proteins were eluted in SDS sample buffer and separated on a 12% SDS-PAGE gel. After transfer to nitrocellulose membranes, immunoblot analysis was performed with RhoA or Rac1 antibody and detected with Femto Western blotting reagent.

Rho and Rac membrane fractionation assay.

Isolated pancreatic acini were treated as noted for 2, 5, 10, and 30 min. Acini were collected into a lysis buffer containing 50 mM HEPES (pH 7.4), 50 mM NaCl, 1 mM MgCl2, 2 mM EDTA, 1 mM PMSF, 10 μg/ml leupeptin, 1 mM Na3VO4, 5 mM NaF, and 1 mM DTT, and lysed by passage five times through a 27-gauge needle, followed by centrifugation for 10 min at 500 g to remove debris and nuclei. The supernatants were then centrifuged at 120,000 g for 45 min in a Beckman Optima TLX ultracentrifuge. After the supernatant was collected as the cytosol fraction, the pellet was resuspended in 1% Triton X-100 in the lysis buffer and centrifuged at 10,000 g for 10 min to remove insoluble debris; the supernatant was collected as the membrane fraction. The amounts of RhoA or Rac1 in each fraction were determined by immunoblot analysis.

Immunolocalization by confocal fluorescence microscopy.

Freshly isolated mouse pancreatic acini were treated with or without CCK (100 pM) for 10 min at 37°C. Acini were allowed to settle in test tubes and fixed for 30 min at room temperature with 4% paraformaldehyde in phosphate-buffered saline, pH 7.4. Acinar preparations were rinsed in phosphate-buffered saline, cryoprotected, and frozen with isopentane cooled with liquid nitrogen. Cryostat sections (5 μm thick) were mounted on SuperFrost Plus slides (Fischer Scientific) and processed for immunofluorescence localization. Briefly, the sections were incubated with primary antibody rabbit anti-Rho or mouse anti-Rac (diluted 1:200) for 2 h at room temperature and washed with PBS three times. The sections were then incubated with secondary antibodies Cy3-conjugated donkey anti-rabbit IgG (1:200) or anti-mouse IgG (diluted 1:200) for another hour and washed. For F-actin visualization in double staining, Alexa 488 phalloidin (1:40) was added to the secondary antibody. For only F-actin staining, rhodamine-phalloidin was diluted 1:400 (Molecular Probes, Eugene, OR) and incubated with sections for 1 h at room temperature. Digitized images were generally collected with an Olympus Fluoview 500 confocal microscope as a Z-series (0.5 μm steps). A single image from the Z-stack was then chosen based on the most informative and sharpest fluorescence and processed using Photoshop CS software (Adobe Systems).

Intracellular Ca2+ measurement.

Intracellular Ca2+ was measured as previously described (54). Acini were incubated with 1 μM fura-2 AM at 37°C for 30 min and then washed and resuspended in fresh HR buffer and transferred to a coverslip in a closed chamber mounted on the stage of a Zeiss Axiovert inverted microscope, and continuously superfused at 1 ml/min with 37°C HR buffer alone or HR buffer containing 300 pM CCK. Solution changes were accomplished rapidly by means of a valve attached to an eight-chambered superfusion reservoir. Measurement of intracellular Ca2+ was performed with the use of a digital imaging system and software (Attofluor, Rockville, MD). Under the conditions used, the presence of EGFP in acini did not interfere with the fura 2 signal or calculated intracellular Ca2+ concentration (9).

F-actin quantitation.

Changes in F-actin content were measured as described previously (42). Briefly, acini were fixed for 15 min with 3.7% formaldehyde in PIPES buffer. Acini were then pelleted and resuspended in 0.1% saponin and incubated with 0.7 μM rhodamine-conjugated phalloidin for 60 min in darkness on a rotator. Stained pellets were washed three times with buffer containing saponin, after which the bound Rhodamine-conjugated phalloidin was extracted with methanol. Protein was measured with the Bio-Rad protein assay. The fluorescence of extracts was measured using excitation at 541 nm and emission at 565 nm. The relative F-actin content was calculated as the ratio of the fluorescence emission per microgram of protein divided by the fluorescence emission per microgram of protein of the control sample.

Statistical analysis.

Results are expressed as means ± SE. Multiple comparisons in the dose-response curves were performed using analysis of variance with Dunnett’s test for comparison with control using GraphPad Prism software. The data presented are means ± SE, with P < 0.05 representing significance.


CCK dose-dependently activates Rho and Rac.

To assess whether CCK dose-dependently activated Rho or Rac in mouse pancreatic acini, acini were treated with different concentrations of CCK for 5 min before being collected, lysed, and analyzed by a pull-down assay using a GST-rhotekin-Rho binding domain, which specifically binds GTP-bound Rho and a GST-PAK-PBD domain, which specifically interacts with GTP-bound Rac. Little active Rho or Rac was detected in untreated cells. CCK treatment activated Rho in a dose-dependent manner with a significant increase at 30 pM and a maximal effect at 10 nM, where active Rho was increased 3.2 ± 0.8 fold (Fig. 1, A and B). Rac activity was also concentration-dependently stimulated by CCK with a significant effect at 10 pM CCK and a maximal increase in response to 1 nM CCK (Fig. 2, A and B).

Fig. 1.

Cholecystokinin (CCK) dose-dependently activates Rho and translocates Rho to membranes in mouse pancreatic acini. A: representative autoradiograph of Rho pull-down assay; total Rho is a Western blot of 10 μg of cell lysate. B: densitometric measurements of Rho activation (means ± SE) from 3 independent experiments. Freshly isolated pancreatic acini were treated with indicated concentrations of CCK for 5 min. Cells were then rapidly lysed and active Rho pulled down from 750 μg of lysate with GST-Rhotekin and visualized by Western blot analysis. C: representative image showing Rho translocation. D: densitometric measurements (means ± SE) from 3 independent experiments. Freshly isolated pancreatic acini were treated with 30 pM CCK for the indicated time. Cells were then rapidly collected, lysed, and cytosol separated from membranes by ultracentrifugation. Membrane fraction proteins (30 μg) were loaded for membrane-bound Rho and 10 μg of cytosol fraction were loaded for cytosol Rho. Rho was visualized by Western blot analysis. *P < 0.05; **P < 0.01, compared with no CCK stimulation.

Fig. 2.

CCK dose-dependently activates Rac and translocates Rac to membrane in acini. A: representative autoradiograph of a Rac pull-down experiment; total Rac is a Western blot of 10 μg of cell lysate. B: densitometric measurements of Rac activation (means ± SE) from 3 independent experiments. Freshly isolated pancreatic acini were treated with indicated concentrations of CCK for 5 min. Cells were then rapidly lysed and active Rac pulled down from 400 μg of lysate with glutathione-S-transferase (GST)-p21 binding domain (PBD) and visualized by Western blot analysis. C: representative image showing Rac translocation. D: densitometric measurements (means ± SE) from 3 independently experiments. Freshly isolated pancreatic acini were treated with 30 pM CCK for indicated time. Cells were then rapidly collected, lysed, and the cytosol was separated from the membrane by ultracentrifugation. Membrane fraction (30 μg) proteins were loaded for membrane-bound Rac and 10-μg cytosol fractions were loaded for cytosol Rho. Rac was visualized by Western blot analysis. *P < 0.05; **P < 0.01, compared with no CCK stimulation.

CCK time-dependently translocates Rho and Rac to membrane fraction.

In resting cells, inactive GDP-bound Rho and Rac are mainly located in the cytoplasm, where they form a complex with RhoGDI. Extracellular stimuli promote the dissociation of Rho and Rac from RhoGDI and their translocation to the membrane where Rho and Rac become active and interact with their downstream effectors. Therefore, the translocation of Rho proteins from cytosol to membrane has been used as an alternative indicator of activation. We next studied the time course of Rho and Rac activation by CCK using this technique. CCK (30 pM) was chosen because this is the optimal concentration for amylase secretion in freshly isolated mouse pancreatic acini. Fresh acini were treated with CCK for 2, 5, 10, and 30 min and membrane fractionation was performed. Both Rho (Fig. 1, C and D) and Rac (Fig. 2, C and D) showed translocation to the membrane fraction after CCK treatment as early as 2 or 5 min with a peak at 10 min with a 3.4 ± 0.3- and 2.3 ± 0.6-fold increase for Rho and Rac, respectively. This was followed by a decrease at 30 min. With the use of the same technique, carbachol also stimulated an increase in the fraction of Rho and Rac associated with membrane (data not shown).

To determine which part of the cell active Rho or Rac translocated to after CCK stimulation, we performed immunofluorescence microscopy using antibodies against Rho or Rac, respectively. In the resting state, the majority of Rho and Rac was located diffusely throughout the cell (Fig. 3), although there was some concentration of Rac near the lumen area overlapping with filamentous actin. Ten minutes after CCK (100 pM) stimulation, some of the Rho and Rac had translocated within the apical region; Rho was mainly located in the subapical area around the lumen, whereas Rac showed strong punctate staining in the area where ZGs are mainly concentrated (Fig. 3).

Fig. 3.

CCK stimulation relocates Rho and Rac in acini. Freshly isolated mouse pancreatic acini were stimulated with (C and D) or without (A and B) CCK for 10 min. Cells were pelleted and frozen secretions were cut and stained with anti-Rho (A and C) or anti-Rac (B and D) antibody (red), and F-actin was stained with Alex 488 phalloidin (green). In the resting state, both Rho and Rac are diffusely distributed throughout the cell with some Rac located to the apical region near the lumen, as shown by the yellow staining. Ten minutes after CCK (10 μM) stimulation, Rho and Rac are more concentrated in the apical region with the Rho is mainly located near the lumen (C), whereas Rac is distributed in a punctate manner near the granule area (D). Bar = 20 μm.

Adenoviral-mediated protein expression.

Terminally differentiated pancreatic acini are refractory to traditional gene transfection. To obtain relatively high expression of protein of interest, an adenoviral gene delivery system was therefore employed. We generated adenovirus-expressing dominant negative HA-tagged RhoN19 and RacN17 as well as adenovirus-expressing C. Botulinum C3, an exotoxin that specifically inactivates Rho through ADP-ribosylation at residue asparagine 41 (2). All three of these adenoviruses also expressed EGFP under a separate CMV promoter; an adenovirus expressing EGFP and β-gal was used as a control. Isolated mouse pancreatic acini were cultured with virus for 16 h before they were collected for Western blot analysis. Over 95% of the cells were infected with initial EGFP expression observed after 6–8 h and strong expression after 16 h. Western blot analysis of HA expression in pancreatic acini infected with adenovirus-expressing dominant negatives for 16 h showed a clear titer-dependent expression of protein (Fig. 4A) without an obvious effect on cell morphology (data not shown). We chose 107 pfu/ml as our standard titer for the studies because these conditions give us strong biological effects. Expression of Cdc42N17 was demonstrated by Western blot analysis for Cdc42 because this construct was not epitope tagged (data not shown).

Fig. 4.

Adenoviral-mediated inhibition of Rho and Rac reduces amylase release in isolated mouse pancreatic acinar cells. A: isolated pancreatic acini were cultured for 16 h with indicated titers of specified adenovirus before cells were collected and expression of protein visualized by Western blot analysis with anti-hemaglutinin (HA)-tag antibody. A clear titer-dependent expression of protein is seen. B: isolated mouse acini were cultured for 16 h with adenovirus expressing β-galactosidase (β-gal) control or C3 exotoxin, then collected and resuspended in HEPES-Ringer buffer (HRB) and stimulated with different concentrations of CCK for 30 min. Amylase release was inhibited by C3 toxin across the CCK dose-response curve. C: acini were cultured for 16 h with adenovirus expressing β-gal control, RhoN19, or RacN17 alone, or the combination of RhoN19 and RacN17, followed by stimulation of amylase release for 30 min. Note the strong additive inhibition of secretion when dominant negative Rho and Rac are both present. Data presented in B and C are the means ± SE from 4–5 experiments. *P < 0.05, compared with β-gal control.

C3, RhoN19, and RacN17 inhibit CCK-induced amylase release.

To determine whether Rho and Rac are involved in secretagogue induced amylase release, isolated acinar cells were cultured with 107 pfu/ml adenovirus-expressing C3, dominant negative Rho, or dominant negative Rac for 16–17 h, and then stimulated with CCK at different concentrations for 30 min and the amylase in the medium measured. As reported before, overnight-cultured acini were not as sensitive as freshly isolated acini to CCK stimulation of amylase secretion (9, 18). In fresh acini, the concentration of CCK inducing maximal amylase secretion was 30 pM, whereas in cultured acini it was 100 pM. Moreover, overnight-cultured acini contained more amylase (2.58 ± 0.04 fold, n = 3) when compared with fresh acini by Western blot analysis. The number of ZGs was more than doubled and extended into the basolateral pole when visualized in toludine blue-stained semithin sections. As a result, the cultured acini secreted less compared with fresh acini when secretion was normalized to the total amylase content: cultured acini released ∼6–9% of total after 30 min, whereas fresh acini released 12–18% of the total. The partial loss of cell polarity may also contribute to the relative low amylase release in overnight-cultured acini. The control adenovirus had no effect on either basal or stimulated secretion compared with no viral infection as reported before (9). Neither C3, dominant negative Rho, nor dominant negative Rac significantly decreased basal amylase release. However, C3 exotoxin and RhoN19 (Fig. 4) both reduced CCK-induced amylase secretion across the amylase release dose curve by 24% and 23%, respectively, and RacN17 decreased it by 32%. The characteristic biphasic CCK dose response curve was preserved in all cases. Simultaneous inhibition of Rho and Rac by using the combination of RhoN19 and RacN17 (Fig. 4, B and C), or C3 and RacN17 (data not shown) resulted in a greater reduction of CCK-stimulated amylase secretion by 68% and also reduced basal amylase release. By contrast, expression of dominant negative Cdc42N17 failed to reduce the amylase secretion under the same conditions (data not shown).

Rho and Rac differentially affect the time course of amylase secretion.

Exocytosis is believed to occur in two phases in pancreatic acini: an initial (early) phase, which is completed within 5–10 min of agonist stimulation, and a late (slow, second, or sustained) phase, which is sustained for the duration of hormonal stimulation (28). We next investigated which phase of secretion was affected by dominant negative Rho or Rac mutants. To perform this experiment, we cultured the acini overnight with adenovirus-expressing control or dominant negative mutants of Rho and Rac. Acini were then treated with 30 pM CCK for 0, 5, 10, 20, and 30 min and the medium was collected to determine the amount of amylase release during the specific stimulation time. Our data showed that RhoN19 reduced both the early and late phases by 42.2 ± 0.03% and 39.8 ± 0.1%, respectively (Fig. 5, A and C). In contrast, RacN17 strongly inhibited the late phase secretion by 59.2 ± 4% but minimally affected the early phase (Fig. 5, B and D). These data suggested that Rho and Rac act differently on CCK-stimulated acinar secretion.

Fig. 5.

Rho and Rac affect the time course of amylase release. Isolated acini were cultured for 16 h with adenovirus-expressing β-gal or RhoN19 (A) or β-gal or RacN17 (B). Cells were then collected and resuspended in HRB and stimulated with or without CCK (30 pM) for indicated times. Secretion was induced by 30 pM CCK (with basal secretion subtracted) during the first 5 min and during the period from 20 to 30 min, comparing the effect of β-gal and RhoN19 (C) and β-gal and RacN17 (D). Data presented are means ± SE for 3 independent experiments. *P < 0.05 and **P < 0.01, compared with β-gal control.

RhoN19 and RacN17 inhibit amylase release induced by carbachol.

To determine whether the effects of dominant negative Rho and Rac on secretion was secretagogue specific, we studied their effects on amylase release induced by carbachol, a secretagogue acting via M3 muscarinic cholinergic receptors to stimulate amylase release. The secretory response to carbachol was diminished by dominant negative Rho and Rac to a similar extent to that of CCK without affecting basal secretion (Fig. 6), suggesting that the effects of Rho and Rac are not stimulus specific but occur after receptor activation.

Fig. 6.

Dominant negative Rho and Rac inhibit amylase release stimulated by carbachol. Isolated acini were cultured with adenovirus-expressing β-gal, RhoN19, or RacN17 for 16 h. Cells were then collected and resuspended in HRB and stimulated with different concentrations of carbachol for 30 min. Results shown are means ± SE for 3 independent experiments. *P < 0.05, compared with β-gal control.

C3, RhoN19, and RacN17 do not affect the CCK-stimulated intracellular Ca2+ increase.

Because of the central role of Ca2+ in secretion, we next examined the effects of C3, dominant negative Rho, or Rac on the intracellular Ca2+ increase in response to CCK (Fig. 7). Acini were cultured with adenovirus-expressing β-gal, dominant negative Rho, dominant negative Rac or C3 for 16 h before being loaded with fura-2 AM for 30 min at room temperature. Acini were then stimulated with CCK (300 pM) and intracellular Ca2+ concentration was determined. None of the three affected basal Ca2+ concentration (90–100 nM) or the amplitude of the intracellular Ca2+ response to maximal CCK stimulation (350–400 nM). These results suggest that the inhibition occurred at a late step of stimulus-secretion coupling downstream of the CCK receptor and intracellular Ca2+ mobilization.

Fig. 7.

Dominant negative Rho and Rac and C3 have no effect on CCK stimulated Ca2+ mobilization. Isolated acini were incubated with 107 pfu/ml of virus expressing β-gal, dominant negative RhoN19, RacN17, or C3 for 16 h. Acini were then resuspended in HR buffer and incubated with fura-2 AM for 30 min, after which intracellular calcium concentration was determined with an Autofluor dual-wavelength imaging workstation. The average basal and peak values of intracellular calcium concentration stimulated with 300 pM CCK were shown. Each result presented is the mean ± SE of 3 independent experiments, including 5–10 cells from each of 6–7 acini.

Latrunculin inhibits secretagogue-induced amylase secretion without altering Ca2+ mobilization.

Rho and Rac are both well known to effect cytoskeletal reorganization and actin remodeling has been suggested to play a critical role in secretion. We next tested the hypothesis that Rho and Rac regulated amylase secretion in acini by reorganizing the actin cytoskeleton. To this end, we employed latrunculin, a toxin from the Red Sea sponge Latrunculia magnificans that specifically binds G-actin near the nucleotide binding site, thereby preventing G-actin polymerization and ultimately leading to depolymerization. To evaluate the effects of latrunculin on the total amount of F-actin, acini were incubated with latrunculin (1 μM) for different time before they were fixed, permeabilized, and loaded with rhodamine phalloidin for F-actin labeling. Phalloidin was extracted with methanol and the intensity of fluorescence was measured and normalized to the protein concentration. Latrunculin induced a rapid reduction in the total F-actin content by 28% after 10 min and a slower subsequent decrease, which reached 44% after 2 h of incubation (Fig. 8A). Fluorescence microscopy of rhodamine phalloidin revealed that latrunculin significantly decreased basolateral F-actin with the apical F-actin relatively resistant to latrunculin treatment (Fig. 8B). Two hours after latrunculin incubation, there was still a substantial amount of F-actin left at the apical region. These data suggested that there are two separate pools of F-actin in acini, although the exact component difference between these two pools remains elusive (4). We also studied the effects of latrunculin on amylase secretion. Latrunculin dose-dependently inhibited amylase secretion across the dose-response curve with 0.3 μM, thereby decreasing amylase secretion by 43% and 1 μM by 70% (Fig. 8C). Similar to CCK stimulation, carbachol-induced amylase release was also inhibited by latrunculin (data not shown). To examine whether these drugs affect agonist-stimulated secretion through interfering with secondary messengers like Ca2+, the cells were preincubated with 3 μM latrunculin for 2 h, after which the intracellular Ca2+ response to CCK stimulation was recorded. Latrunculin treatment did not affect the Ca2+ response to CCK stimulation (Fig. 8D), suggesting that latrunculin affects the latter steps in secretion.

Fig. 8.

Latrunculin decreases the total content of F-actin and inhibits amylase release without change in Ca2+ mobilization. A: time course of the total F-actin content after latrunculin treatment. Acini were treated with latrunculin (1 μM) for the indicated time and the F-actin content was measured. B: representative confocal images of F-actin staining showing reduced F-actin after latrunculin (1 μM) treatment for 2 h. C: amylase assay in freshly isolated pancreatic acini preincubated with or without latrunculin A (0.3 μM and 1 μM) for 30 min before CCK stimulation for 30 min. D: average basal and peak values of intracellular calcium concentration stimulated with 300 pM CCK in fresh acini pretreated with 3 μM latrunculin for 2 h. Each result presented is the mean ± SE of 3 independent experiments.


The Rho family of small G proteins is well documented as a regulator of the actin cytoskeleton. However, their effects on secretion in pancreatic acinar cells are incompletely understood. In the present study, we demonstrated for the first time that Rho and Rac are both activated by CCK in pancreatic acinar cells, and that after activation, both of them translocate within the apical region where exocytosis occurs. We also provide evidence that the activation of Rho and Rac is required for stimulated acinar secretion. Rho and Rac may well have different functions in secretion as dominant negative Rho inhibited both the early and late phases of amylase secretion, whereas dominant negative Rac inhibited mainly the late phase and the two were additive in their overall effect. Inhibition of secretion occurred without an effect on the secretagogue-induced increase in intracellular Ca2+ and was mimicked by the actin depolymerizer latrunculin. Together, these observations provide direct evidence for the involvement of Rho and Rac in pancreatic acinar cell secretion and thus provide a new signal pathway in pancreatic acini that regulates secretion besides the traditional calcium mobilization. Along with other recent data (30), this study suggested that this involvement of Rho and Rac on secretion may be mediated by the actin cytoskeletal reorganization.

In pancreatic acini, the terminal web under the apical membrane has been proposed to act both positively and negatively on secretion. Manipulation of actin dynamics with pharmacological agents has been shown to be sufficient to trigger or block exocytosis (7, 28, 49). Recently, actin has been shown to also coat ZGs in the course of exocytosis. One study reported occasional actin-coated ZGs, which where interpreted as occurring before exocytosis and having a short life span (47). By contrast, a recent study by Nemoto et al. (30), using two photon confocal microscopy for direct visualization of ZG exocytosis in pancreatic acini, demonstrated that physiological concentrations of CCK induced F-actin coating of all granules undergoing exocytosis and showed that the lifetime of these actin-coated ZGs was relatively long. In this latter study (30), latrunculin (10 μM) was shown to decrease F-actin at the apical region and prevent F-actin coating of ZGs. Acini incubated with C3 exotoxin for 2 h showed minimal effects on the apical F-actin content, whereas 3-day incubation almost totally ablated apical F-actin (30), suggesting that Rho is involved in the apical actin polymerization and Rho-dependent actin polymerization occurs at the apical membrane at a slow rate. In addition to the blockage of actin coating on ZGs, C3 and latrunculin were also shown to induce large spherical intracellular vacuoles with a diameter of >2 μm, when cells were stimulated with 100 pM CCK at room temperature. Their observation was that these vacuoles lack F-actin coating and seemed to form by the merger of fused granules and expanded with time during the secretagogue stimulation. They interpreted these observations as indicating that the function of F-actin coating was to prevent vacuole formation by stabilizing the granules undergoing exocytosis. Nemoto et al. (30) concluded that Rho-dependent F-actin coating of granule membranes stabilizes exocytotic structures and is necessary for physiological progression of sequential compound exocytosis in the exocrine pancreas and for prevention of acute pancreatitis. In our experiments, we preincubated the acini with 1 μM latrunculin and found that it decreased the total F-actin content quickly by 28% at 10 min, followed by a slower decrease without obvious cellular morphological changes. The apical actin web appeared more resistant to latrunculin treatment compared with the basolateral actin although the compositional and functional differences between these two pools are unclear. Preincubation of acini with latrunculin or overexpression of C3 by means of adenoviral infection decreased the secretory response to both CCK and carbachol comparably. Although RhoN19, RacN17, and C3 had minimal effects on the total F-actin content after 16 h (data not shown), we concluded that the main action of C3 and latrunculin was to prevent actin polymerization. This suggests that it is the local actin rearrangement especially at the apical region and not the amount of total F-actin that determines or affects the process of secretion and also suggests that actin polymerization plays a positive role in secretion. Latrunculin and C3 may also affect secretion by affecting the actin coating of ZGs as C3 and latrunculin were shown to prevent the coating of ZGs after secretagogue stimulation (30). Whether dominant negative Rho and Rac may affect the actin coating of ZGs to impact secretion has not been studied. Another study (41) addressed the function of Rho in pancreatic acinar secretion by incubating C3 exotoxin (25 μg/ml) with isolated rat pancreatic acini for 2 h and showed that C3 partially blocked amylase release stimulated by CCK. Our studies used an adenoviral gene delivery system to express C3 exotoxin and dominant negative mutant of Rho in mouse pancreatic acini and showed that expression of both C3 and RhoN19 decreased secretory response to both CCK and carbachol. Our data clarify the previous observation and suggest a positive role of Rho in pancreatic acinar cell secretion.

However, the finding by Nemoto et al. (30) that C3 and latrunculin both induced intracellular vacuoles raised the possibility that C3 and latrunculin may not influence secretion per se but rather stimulate the formation of vacuoles, sequester digestive enzymes inside them, and thereby inhibit secretion. To test this possibility, we examined the vacuole formation in our acinar cells. Several major differences in experimental conditions were noticed. First, the concentration of latrunculin was different. In our study, we used a concentration of 1 μM, whereas Nemoto et al. (30) used a 10 μM concentration. Although no other binding partners have been identified for latrunculin, it is still possible that high concentrations of pharmacological agents may have nonspecific effects. Second, all of our experiments were carried out at 37°C, whereas those of Nemoto et al. were at a room temperature that varied between 22° and 25°C. Therefore, to determine the influence of temperature on acinar physiological functions and morphological changes, we carried out a series of experiments. At room temperature, the amylase release curve was different from that obtained at 37°C in that maximal amylase release was reduced without showing high-dose inhibition. Also, toludine blue staining of the semithin section of pancreatic acini showed that temperature also played an important role in the vacuole formation. Under conditions similar to those in Nemoto study, stimulation of 100 pM CCK for 10 min at room temperature induced some vacuoles, so did 10 μM latrunculin. The combination of latrunculin and CCK evoked more and larger vacuoles as shown in Nemoto’s study. However, under the conditions of our experiments at 37°C, neither 1 μM latrunculin, nor 30 pM CCK, or the combination of latrunculin and CCK had major effects on intracellular vacuole formation (data not shown). Thus temperature is an important factor that can affect experimental results. In addition, we examined the vacuole formation in acini cultured with adenovirus-expressing C3 exotoxin and failed to detect any pathological vacuole formation. Therefore, the inhibition of amylase release caused by C3 or dominant negative Rho is not due to sequestration of amylase within vacuoles under the conditions at which we performed our experiments, although it may well occur at room temperature.

We also addressed the function of Rac and Cdc42 in pancreatic acinar cell secretion. To our knowledge, no previously studies have evaluated their role in pancreas. With the use of an adenoviral gene delivery system to express dominant negative mutants of Rac and Cdc42 in mouse pancreatic acini, our results indicated that Rac also plays a positive role in secretagogue-induced amylase release as an expression of dominant negative Rac inhibited amylase release across the dose curve stimulated by CCK and carbachol. This is in good agreement with studies in other cell systems, including neurons (32), permeabilized mast cells (6, 23, 37), and bovine chromaffin cells (24), where a role for Rac in secretion has been observed. Although some reports (6, 21) have related Cdc42 positively with secretion in mast cells and pancreatic β-cells, dominant negative Cdc42N17 failed to inhibit secretion stimulated by secretagogues in pancreatic acini although Cdc42 was activated by CCK in acini as shown by a pull-down assay with GST-PAK (data not shown).

We also evaluated the effects of Rho and Rac on the time course of the secretagogue-induced secretion. In pancreatic acini, exocytosis has been proposed to consist of an initial early phase and a late phase (51, 53). The early phase has been related to intracellular Ca2+ mobilization or the fusion of docked granules in the vicinity of the plasma membrane, whereas the late phase has been related to Ca2+ influx or release from a reserve pool of granules. In the present study, dominant negative Rho decreased both the early and late phases of secretion, whereas dominant negative Rac only suppressed the late phase. The effect of Rac on the late phase of secretion was consistent with another study in INS-1β cells (23). Because the intracellular calcium response to CCK stimulation was intact in C3-, RhoN19-, and RacN17-infected acini, Rho and Rac are unlikely to affect the early steps in receptor mediated signaling but rather are more likely involved in the final steps of pancreatic acinar exocytosis, including granule transport, docking, priming, or fusion. Two reports (11, 23) suggested that Rac regulates calcium-dependent exocytosis in other cell types, most likely through an effect on the number of readily releasable vesicles. Interestingly, Rac also appeared to localize differently from Rho at the apical region visualized by confocal immunofluorescence.

Rho and Rac seem to regulate amylase secretion independently as inhibition of both Rho and Rac resulted in a greater reduction of secretion than inhibiting either one alone; second, Rho decreased both early and late phases of secretion, whereas Rac only suppressed the late phase. The mechanisms underlying the different roles of Rho and Rac in acinar secretion are poorly understood. Different localization of Rho and Rac may partially explain the difference. In acini, the majority of Rho and Rac was located throughout the cell although there was some Rac near the lumen area in the resting state, and after CCK stimulation Rho was mainly located in the subapical area around the lumen, whereas Rac showed strong punctuate staining in the area where ZGs are mainly concentrated. Because it is clear that both Rho and Rac are major factors regulating actin cytoskeleton in many cells (8) and dominant negative Rac only reduced the late phase of secretion and Rac localized in the granule area, it seems likely that Rac promotes granule movement to release sites, whereas Rho may regulate actin polymerization around granules already in a releasable pool as suggested by Nemoto et al. (30).

Unfortunately, the downstream effectors of Rho and Rac that regulate secretion remain obscure. Rho kinase was the first effector of Rho to be discovered and is the best characterized for its role in mediating the formation of stress fibers and focal adhesions through the inhibition of myosin light chain phosphatase, which correlates with increases myosin light chain phosphorylation and secretion in several cell types (25, 33). In pancreatic acini, myosin I was found to be associated with ZG membranes, and myosin II was found to be distributed both apically and underneath the basolateral plasma membrane (35). Inhibition of myosin II with 2,3-butanedione monoxime, a well-characterized inhibitor of myosin II ATPase activity, or inhibiting myosin light chain phosphorylation with 1-(5-chlornaphthalene-1-sulfonyl)-1H-hexahydro-1,4-diazepine, a myosin light chain kinase-specific protein kinase inhibitor, both completely inhibited CCK-induced amylase release (46). However, the Rho kinase inhibitor Y-27632 (10 μM) had no influence on either basal or CCK-induced amylase secretion suggesting that Rho kinase is not necessary for amylase release (Y. Bi and J. A. Williams, unpublished data). A lack of effect of Y-27632 was also reported in mast cells (43) and rabbit parietal cells (45) in both of which Rho was implicated whereas Y-27632 did not have any effect on secretion, although one study in PC12 cells reported that Rho kinase partially mediated Rho action on secretion (13). Other Rho effectors, such as phosphatidylinositol 4-kinase, might be involved in secretion by promoting the formation of granule-associated actin filaments required for late stages of exocytosis because phosphoinositides have been reported to be associated with secretory granules (14), initiate actin nucleation, and regulate membrane-cytoskeleton interactions, and been suggested to mediate the Rho effects on secretion in chromaffin cells (14, 15). Multiple Rac effectors have also been found implicated in inducing actin polymerization (10) and may regulate secretion by playing a role in ZG transport; however, it is not clear which effector is responsible for secretion in acini. Therefore, the downstream effectors mediating the Rho and Rac on secretion in pancreatic acinar cells remain elusive. Further investigations will have to determine the precise signaling cascades elicited by Rho and Rac, which contribute to the regulation of secretion.


This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-41122 and DK-52860 (to J. A. Williams), the Michigan Gastrointestinal Peptide Center Grant P30 DK-34933, and the Michigan Diabetes Research and Training Center Grant P60 DK-20572.


We thank Sophie Le Page for assistance with preliminary studies, Weizhen Zhang for technical assistance in intracellular calcium measurement, Baoan Ji for critical discussion and technical support on virus construction, Brad Nelson for embedding and sectioning the pancreatic acini, and Stephen Lentz for help with confocal microscopy.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract