Why Mouse Airway Submucosal Gland Serous Cells Do Not Secrete Fluid in Response to cAMP Stimulation*

Background: Mouse submucosal glands exhibit lower rates of cAMP-activated CFTR-dependent fluid secretion than human glands. Results: Mouse serous cells exhibit CFTR chloride permeability but lack cAMP-activated calcium signals required for counterion permeability and secretion. Conclusion: Lack of cAMP-activated calcium signaling accounts for weak mouse serous cell fluid secretion. Significance: This result explains a fundamental difference between human and mouse airway glands. Airway submucosal glands are important sites of cystic fibrosis transmembrane conductance regulator (CFTR) chloride (Cl−) channel expression and fluid secretion in the airway. Whereas both mouse and human submucosal glands and their serous acinar cells express CFTR, human glands and serous cells secrete much more robustly than mouse cells/glands in response to cAMP-generating agonists such as forskolin and vasoactive intestinal peptide. In this study, we examined mouse and human serous acinar cells to explain this difference and reveal further insights into the mechanisms of serous cell secretion. We found that mouse serous cells possess a robust cAMP-activated CFTR-dependent Cl− permeability, but they lack cAMP-activated calcium (Ca2+) signaling observed in human cells. Similar to human cells, basal K+ conductance is extremely small in mouse acinar cells. Lack of cAMP-activated Ca2+ signaling in mouse cells results in the absence of K+ conductances required for secretion. However, cAMP activates CFTR-dependent fluid secretion during low-level cholinergic stimulation that fails to activate secretion on its own. Robust CFTR-dependent fluid secretion was also observed when cAMP stimulation was combined with direct pharmacological activation of epithelial K+ channels with 1-ethyl-2-benzimidazolinone (EBIO). Our data suggest that mouse serous cells lack cAMP-mediated Ca2+ signaling to activate basolateral membrane K+ conductance, resulting in weak cAMP-driven serous cell fluid secretion, providing the likely explanation for reduced cAMP-driven secretion observed in mouse compared with human glands.

Airway submucosal exocrine glands are major sites of fluid secretion in the lung and are likely important in the pathology of cystic fibrosis (CF) 2 (1)(2)(3). CF is caused by defects in the function of the cystic fibrosis transmembrane conductance reg-ulator (CFTR) anion channel (4), which is expressed both in the surface airway epithelium and in the serous acinar cells at the distal ends of submucosal glands. Fluid secretion in response to intracellular cAMP ([cAMP] i )-elevating agonists is severely reduced in glands from CF patients (5)(6)(7), CFTR-knock-out pigs (8), and cftr tm1UncϪ/Ϫ knock-out (cftr Ϫ/Ϫ ) mice (9,10). We previously observed that agonist-evoked elevations of cAMP in porcine and human serous acinar cells evoke protein kinase A (PKA)-mediated fluid secretion that is dependent upon CFTR, which functions as an apical membrane secretory Cl Ϫ channel (11). Defective cAMP-activated CFTR-dependent serous cell fluid secretion likely contributes to airway dehydration and/or altered mucus rheology seen in CF lung disease (1)(2)(3). Restoration of this fluid secretion may be an important therapeutic strategy for CF, and thus the elucidation of the molecular mechanisms of serous cell secretion is critically important.
Fluid secretion by exocrine acinar cells is driven osmotically by the primary secretion of NaCl (reviewed in Ref. 12). Cl Ϫ is secreted through the cell by uptake mechanisms in the basolateral membrane operating in series with a Cl Ϫ channel in the apical membrane that mediates its efflux. The resulting transepithelial voltage drives Na ϩ from the serosal side into the gland lumen via paracellular pathways. Regulation of apical membrane secretory Cl Ϫ channels is a key step in the activation of fluid secretion. However, secretory cells also require sufficient counter-ion permeabilities to preserve electroneutrality during the robust efflux of cellular anion content that drives fluid secretion. The counter-ion permeability is typically provided by K ϩ channels localized on the basolateral membrane. A sufficiently high K ϩ conductance is necessary to maintain membrane potential (V m ) sufficiently hyperpolarized to provide an electrical driving force for Cl Ϫ secretion (11)(12)(13)(14)(15). Although the apical membrane Cl Ϫ conductance is usually rate-limiting for activation of fluid secretion in most exocrine cells, we previously discovered that activation of CFTR was insufficient to stimulate fluid secretion in human and porcine submucosal gland serous acinar cells because the basal K ϩ conductance was small and insufficient (11). Nevertheless, cAMP activated CFTR-dependent fluid secretion in porcine bronchial and human nasal serous acinar cells because it also elicited cytoplasmic Ca 2ϩ ([Ca 2ϩ ] i ) signals that activated K ϩ channels to provide the necessary counter-ion permeability for fluid secretion (11). In contrast with human and pig serous acinar cells, cAMP failed to activate fluid secretion in mouse serous acinar cells (15). These observations were in accord with data demonstrating that intact mouse submucosal glands exhibit much lower rates of cAMP-activated, CFTR-dependent fluid secretion (normalized to the maximum rate of cholinergic-activated fluid secretion) compared with either human or porcine submucosal glands (3, 5-7, 9, 10, 16 -19). Because we have shown that mouse, porcine, and human serous cells have similar apical membrane CFTR staining (11,14,15,20), we hypothesized that, like human and pig serous acinar cells, mouse acinar cells lack a resting K ϩ conductance, but mouse cells differ in their ability to secrete fluid in response to [cAMP] i -elevating agonists because of a lack of cAMP-activated robust K ϩ permeability, potentially due to the absence of cAMP-activated Ca 2ϩ signaling.
Here, we have tested this hypothesis as well more thoroughly examined the role of CFTR in fluid secretion by mouse serous acinar cells. We utilized optical methods developed in rat salivary gland acinar cells (13,21,22) and adapted for mouse, human, and porcine airway submucosal gland serous acinar cells (11,14,15,20) to measure fluid secretion in freshly isolated intact living cells. Fluorescence imaging of Ca 2ϩ and Cl Ϫ was combined with simultaneous differential interference contrast (DIC) imaging of cell volume that reflect changes in the secretory state of the cells (11,14,15,20). Our results suggest that while human and mouse serous acinar cells share many of the mechanisms that drive fluid secretion, fundamental differences exist in the secretagogue-induced second messenger pathways that activate secretion. These data have important implications for explaining differences in the rates of cAMP-induced fluid secretion observed between mouse and human submucosal glands. In addition, they provide insights into the utility of mouse serous cells as a model for CFTR-dependent fluid secretion.
All solutions used were exactly as described in Refs Ϫ and contained 20 mM HEPES, 2 mM L-glutamine, 1ϫ penicillin/streptomycin, 50 g ml Ϫ1 gentamycin, and 1ϫ MEM vitamins, amino acids, and nonessential amino acids. Collagenase digestion was performed in Solution B lacking added CaCl 2 but with no EGTA added. The control solution for NO 3 Ϫ substitution experiments (Solution C (11,23) Isolation and Imaging of Primary Serous Acinar Cells-All procedures were carried out exactly as described (11,14,15,20). Mouse tissues were obtained from animals that were euthanized by CO 2 inhalation for unrelated experiments. No animals were killed specifically for these studies. Human tissue was obtained with written patient consent from surgical specimens collected after turbinectomy procedures performed at the Rhinology Clinic of the Hospital of the University of Pennsylvania. Animal and human tissues were obtained with full approval of the University of Pennsylvania Institutional Animal Care and Use Committee (IACUC) and Institutional Review Board (IRB), respectively. Minced mouse nasal turbinate and septum tissue or dissected human submucosal glands were incubated in Ca 2ϩ -free Solution B containing 1.5 mg ml Ϫ1 Type IV collagenase and 10 g ml Ϫ1 DNase I for ϳ40 min at room temperature with gentle shaking. Following three washes (1000 RPM for 1 min in a clinical centrifuge) with collagenase free-Solution B (containing Ca 2ϩ ), cells were resuspended in Solution A with 95% O 2 /5% CO 2 gassing and plated on washed, uncoated glass coverslips.
CFTR and NKCC1 immunofluorescence staining was performed as previously described (11,14,15,20). Isolated acini and cells were plated on Cell-Tak coated coverslips and fixed for 20 min in 4% formaldehyde at room temperature. Blocking and permeabilization was performed for 1 h in PBS containing 1% BSA, 2% normal goat serum, and 0.15% saponin. Primary antibody (anti-NKCC1 at 1:400 and/or anti-CFTR at 1:100) incubation was performed overnight at 4°C. Secondary antibody incubation (both anti-rabbit and/or anti-mouse at 1:1000) was performed for 2 h at 4°C. Coverslips containing cells were mounted on slides using Vectashield Hard-Set mounting medium and viewed using a Perkin-Elmer Ultraview spinning disk confocal system and 488 nm and 568 nm laser lines as previously described (11,14,15,20).
Statistics and Data Analysis-Images were acquired using Ultraview Software (PerkinElmer Waltham, MA). Data were analyzed using Excel, Igor Pro (Wavemetrics, Inc., Lake Oswego, OR) and/or ImageJ (W. Rasband, NIH, Bethesda, MD). All values are reported as means Ϯ S.E. of the mean (S.E.). Student's t test and one-way analysis of variance (ANOVA) with Bonferroni or Dunnett's multiple comparison post-tests were used to determine statistical significance (p values) as appropriate; a p value of Ͻ0.05 was considered statistically significant. For all figures, one asterisk (*) indicates p Ͻ 0.05, two asterisks (**) indicate p Ͻ 0.01, and n.s. indicates no statistical significance.  (11)). As shown previously, shrinkage is caused by agonist-induced loss of cell KCl, through CFTR and K ϩ channels, and osmotically obliged water and reflects activation of fluid secretion (11,14,15,20). In contrast, we observed no detectible shrinkage during forskolin or VIP stimulation of wild-type (Wt) mouse serous cells ( Fig. 1C; n ϭ 4; and previously in Ref. 15). However, the mouse cells were viable and capable of secreting Cl Ϫ and fluid because subsequent exposure to 10 M carbachol (CCh; a cholinergic agonist) triggered rapid and substantial cell shrinkage (Fig. 1C). Strong cholinergic stimulation activates fluid secretion by triggering a substantial [Ca 2ϩ ] i rise that activates Ca 2ϩ -activated Cl Ϫ channels (CaCCs; likely including Ano1 (14)), bypassing a requirement for CFTR in mouse cells as well as in human and pig serous acinar cells (11,14,15). Cl Ϫ accumulation in mouse serous cells is overwhelmingly dependent on the activity of the Na ϩ -K ϩ -2Cl Ϫ co-transporter isoform 1 (NKCC1) (15). To determine whether the failure to observe cAMP-induced shrinkage of mouse acinar cells was due to compensatory solute uptake, forskolin stimulation was performed in the presence of the NKCC1 inhibitor bumetanide (100 M). Nevertheless, no forskolin-activated shrinkage was observed under these conditions ( Fig. 1D; n ϭ 4), despite intact secretory responses to CCh. Thus, cAMP fails to activate KCl loss from mouse serous acinar cells, in contrast to its effects in human cells.

Stimulation of Mouse Serous Cells With [cAMP] i -elevating Agonists Activates a PKA-and CFTR-dependent Plasma Membrane Anion
Permeability-The forskolin-and VIP-induced secretory responses of human serous cells were inhibited by the CFTR inhibitor CFTR inh 172 ( Fig. 1, E and F; n ϭ 4 each), demonstrating that CFTR functions as a secretory Cl Ϫ channel during human acinar cell fluid secretion. The lack of forskolininduced secretory responses in mouse cells could be caused by lack of a CFTR-dependent Cl Ϫ conductance. Although this seemed unlikely based on strong CFTR immunofluorescence signals observed in mouse submucosal gland serous cells ( Fig. 2 and Ref. 15,20), it is possible that CFTR, while highly expressed in these cells, may nevertheless be nonfunctional. To determine if CFTR is functional in the plasma membranes of mouse serous acinar cells, we employed a direct and sensitive method to measure plasma membrane Cl Ϫ permeability. Mouse nasal serous acinar cells were loaded with the Cl Ϫ -sensitive fluorophore SPQ, previously used to track Cl Ϫ secretion in these cells (15), and studied using a NO 3 Ϫ substitution protocol (23) that was previously employed in pig and human serous cells (11). SPQ is quenched by Cl Ϫ but not by NO 3 Ϫ . Because Cl Ϫ channels have a high intrinsic permeability to NO 3 Ϫ , replacement of bath Cl Ϫ with NO 3 Ϫ results in electroneutral exchange of intracellular Cl Ϫ for extracellular NO 3 Ϫ , resulting in an increase in SPQ fluorescence that can be used as a measure of relative cellular anion permeability. All experiments were performed in the presence of 50 M bumetanide and in the absence of extracellular CO 2 /HCO 3 Ϫ (0-CO 2 /HCO 3 Ϫ ) to minimize non-conductive Cl Ϫ efflux through NKCC1 and Cl Ϫ /HCO 3 Ϫ (anion) exchanger (AE) activity, respectively.
In unstimulated serous acinar cells from Wt and cftr Ϫ/Ϫ mice, replacement of bath Cl Ϫ with NO 3 Ϫ resulted in a slow increase in normalized SPQ fluorescence ( Fig. 3A; ⌬F/F (t ϭ 0) ϭ 0.0011 Ϯ 0.0004 units⅐s Ϫ1 in Wt and 0.0012 Ϯ 0.0003 units⅐sec Ϫ1 in cftr Ϫ/Ϫ ; n.s.; n ϭ 4 each) reflecting a small basal Cl Ϫ permeability. However, following stimulation of the same cells with 5 M forskolin, introduction of NO 3 Ϫ caused a ϳ10fold more rapid increase in SPQ fluorescence in Wt cells (⌬F/F (t ϭ 0) ϭ 0.013 Ϯ 0.02 units⅐s Ϫ1 ) but not in cftr Ϫ/Ϫ cells (0.0013 Ϯ 0.0004 units⅐sec Ϫ1 ; p Ͻ 0.01 versus Wt). Thus, forskolin activates a CFTR-dependent plasma membrane anion permeability in mouse acinar cells, likely CFTR itself. To confirm this and to ensure that the responses in the Wt cells were not due to activation of CaCC, similar experiments were performed in Wt (average traces shown in Fig. 3B), cftr Ϫ/Ϫ , and cftr ϩ/Ϫ heterozygote (het) serous cells (average traces shown in Fig. 3, B-D, respectively) under similar 0-CO 2 /HCO 3 Ϫ /bumetanide conditions with the cells loaded with the Ca 2ϩ chelator BATPA-AM (11,14,15,20) and stimulated in the absence of extracellular Ca 2ϩ (ϩ1 mM EGTA; 0-[Ca 2ϩ ] o ). SPQ fluorescence changes were recorded upon introduction of NO 3 Ϫ in unstimulated cells or cells stimulated for 180 -200 s with either 5 M forskolin, 3 M VIP, or the ␤-adrenergic agonist isoproterenol (10 M). Basal Cl Ϫ permeabilities were identical in unstimulated Wt, cftr Ϫ/Ϫ , and het cells (Fig. 3E). In the Wt cells, anion permeability was increased ϳ10-fold after stimulation with forskolin, VIP, or isoproterenol (Fig. 3E). The forskolin-activated increase in anion permeability was abolished by 10 M H89 (PKA inhibitor) or 12 M CFTR inh 172 (Fig. 3E). Again, no significant increase in anion permeability was observed after stimulation of cftr Ϫ/Ϫ cells with forskolin, VIP, or isoproterenol (Fig. 3E). Het cells also exhibited a robust forskolin-induced increase in anion permeability (Fig. 3E). While the forskolin-induced permeability in het cells initially appeared smaller than the permeability in Wt cells, the two responses were not significantly different. These results demonstrate that Wt and het mouse cells respond to elevations of [cAMP] i with robust CFTR activation, and suggest that the lack of significant cAMP-activated fluid secretion is not caused by insufficient Cl Ϫ permeability.

Mouse Serous Acinar Cells Lack cAMP-activated [Ca 2ϩ ] i Signals That Are Correlated With and Required for Optimal
Secretion-Acinar cell shrinkage during secretagogue stimulation requires both Cl Ϫ efflux through secretory Cl Ϫ channels as well as counter-ion K ϩ efflux (typically across the basolateral membrane) to preserve cellular electroneutrality. Since the mouse cells possess cAMP-activated CFTR Cl Ϫ permeability, we hypothesized that the lack of cAMP-induced fluid secretion by mouse serous cells is due to a lack of sufficient K ϩ conductance. In human serous acinar cells, cAMP-evoked cell shrinkage requires not only PKA-mediated activation of CFTR-de- showing lack of CFTR immunofluorescence. NKCC1 co-staining was performed as a control for the staining protocol as well as to illustrate the pattern of basolateral membrane staining. pendent anion permeability, but also PKA-activated elevation of [Ca 2ϩ ] i (shown in Fig. 4, A and B) required to activate K ϩ conductance (11), likely mediated by Ca 2ϩ -activated K ϩ channels. Because forskolin failed to cause shrinkage of mouse serous cell despite activation of a robust anion permeability, we speculated that basal K ϩ conductance was low in mouse acinar cells, and that lack of cAMP-activated [Ca 2ϩ ] i signaling might account for the failure of K ϩ conductance activation during stimulation. We therefore examined whether mouse cells exhibit forskolin-induced [Ca 2ϩ ] i signals as observed in pig and human acinar cells (11). Ratiometric imaging of [Ca 2ϩ ] i with fura-2 revealed that neither forskolin nor VIP altered [Ca 2ϩ ] i (Fig. 4, C and D; n ϭ 6). In contrast, 10 M CCh activated a robust elevation of [Ca 2ϩ ] i that caused marked secretion (Fig. 4, C and D) likely mediated by CaCC (15). These data suggest that the low rate of cAMP-activated fluid secretion from mouse serous cells may be caused by lack of cAMP-dependent [Ca 2ϩ ] i signals to activate counter-ion K ϩ conductance.
Forskolin Potentiates Secretory Responses of Mouse Serous Cells to a Sub-secretory [CCh] Through a PKA-and CFTR-dependent Mechanism-Whereas cAMP stimulation results in minimal fluid secretion from mouse serous cells, we asked whether their CFTR anion permeability nevertheless plays a role in fluid secretion under particular conditions. Specifically, we examined the fluid secretion responses to elevated cAMP during weak cholinergic stimulation that moderately raises [Ca 2ϩ ] i . In Wt cells, stimulation with 100 nM CCh elicited a small [Ca 2ϩ ] i signal (147 Ϯ 6 nM; n ϭ 6) that was associated with a minor shrinkage response (Ͻ5% volume change; Fig. 5A). However, after stimulation of the same cells with 5 M forskolin, which alone elicited no [Ca 2ϩ ] i or cell volume responses (Figs. 1 and 4), subsequent stimulation with 100 nM CCh (in the continued presence of forskolin) elicited robust cell shrinkage (15 Ϯ 2%; Fig. 5A). Of note, forskolin had no effect on the magnitude of the 100 nM CCh-evoked [Ca 2ϩ ] i elevation (Fig. 5G), in contrast to the potentiation observed in human serous acinar cells (11). The forskolin-induced enhancement of 100 nM CChevoked cell shrinkage was abolished by the PKA-inhibitor H89 (Fig. 5B), suggesting that forskolin potentiated the secretory response by activating CFTR. In agreement, the potentiated secretory response to 100 nM CCh after 5 M forskolin stimulation observed in Wt cells was absent (Ͻ5% shrinkage) in cftr Ϫ/Ϫ cells, being similar to that observed during stimulation with 100 nM CCh in the absence of forskolin (Fig. 5, C and D). These data suggest that, in mouse serous cells, cAMP potentiates the secretory responses to low-level cholinergic stimulation solely via a PKA-and CFTR-dependent mechanism. This is in contrast to the CFTR-independent secretion observed during combined strong cAMP and weak cholinergic stimulation in human and porcine serous acinar cells (11). Neither forskolin (Fig. 5, E and F) nor the absence of CFTR (Fig. 5G) had an effect on peak [Ca 2ϩ ] i or magnitude of cell shrinkage in response to strong cholinergic stimulation (1 M CCh), as previously observed (15). These results suggest that Cl Ϫ efflux is not ratelimiting during strong cholinergic stimulation, as activation of CFTR-dependent anion permeability does not detectibly enhance shrinkage. Furthermore, they confirm previous observations that CFTR plays little role in secretion during strong cholinergic stimulation (15).
The K ϩ Channel Activator EBIO Stimulates CFTR-dependent Fluid Secretion by Forskolin-stimulated Mouse Serous Acinar Cells-We hypothesized that the synergistic activation of fluid secretion by forskolin and low [CCh] was mediated by cAMP activation of CFTR and activation of basolateral membrane K ϩ conductance by the CCh induced rise of [Ca 2ϩ ] i . To more directly test whether K ϩ conductance is indeed rate-lim-  iting during cAMP-agonist stimulation of mouse serous cells, we used EBIO to directly activate epithelial Ca 2ϩ -sensitive K ϩ channels (24 -27). EBIO stimulation of Cl Ϫ secretion in T84 cell monolayers, rat colonic epithelium, and mouse tracheal epithelium has an EC 50 of ϳ500 M (25,26). A low [EBIO] (100 M) was used in this study as higher levels (Ͼ200 M) can elevate [cAMP] i in isolated mouse colonic crypts (24,27). Nevertheless, we tested whether EBIO elevated [cAMP] i in serous acinar cells by performing NO 3 Ϫ substitution experiments in SPQ-loaded serous acinar cells prior to and during stimulation with 100 M EBIO. EBIO alone did not affect the rate of SPQ fluorescence change upon introduction of NO 3 Ϫ (Fig. 6A), suggesting that at 100 M it does not elevate [cAMP] i sufficiently to activate CFTR. When applied alone, EBIO had no effect on serous acinar cell volume (Fig. 6, A and B) or [Ca 2ϩ ] i (Fig. 6B). However, EBIO caused robust cell shrinkage of forskolin-stimulated Wt cells without affecting [Ca 2ϩ ] i (Fig. 6C). Importantly, EBIO was without effect in forskolin-stimulated cftr Ϫ/Ϫ cells (Fig. 6D). Thus, EBIO synergizes with forskolin to activate secretion in a manner highly reminiscent of the effects of lowlevel cholinergic stimulation, suggesting that they work by similar mechanisms. In agreement, EBIO did not enhance secre- tion induced by 100 nM CCh (Fig. 6E). These results suggest that Cl Ϫ conductance is rate-limiting during low-level cholinergic stimulation. Furthermore, our data suggest that forskolin fails to activate robust secretion in mouse serous acinar cells because of lack of sufficient K ϩ conductance. Based on the above data and (15,20), a model of the molecular mechanisms involved in mouse serous cell fluid secretion is outlined in Fig. 7.

DISCUSSION
While much has been learned about CFTR function from cftr Ϫ/Ϫ transgenic mice, the lack of significant lung pathology in cftr Ϫ/Ϫ knock-out mice has hampered attempts to understand the development CF lung disease (28). The more recent development of transgenic cftr Ϫ/Ϫ pigs (29 -32) and ferrets (33)(34)(35) promises to provide animal models of CF that exhibit more human-like disease phenotypes. However, because of the time and costs required to breed, characterize, and distribute these models, transgenic cftr Ϫ/Ϫ and mutant mice will likely remain important models in which to study CFTR function. Because of this, we examined mouse serous acinar cells in more detail to determine whether or not we could detect CFTR function, utilizing Wt and cftr tm1UncϪ/Ϫ mice (28), with an emphasis on the potential contributions of CFTR to fluid secretion.  A and B). SPQ ⌬F/F (t ϭ 0) was 0.0012 Ϯ 0.0002 units⅐s Ϫ1 (n ϭ 5) before EBIO application and 0.0013 Ϯ 0.0002 units⅐s Ϫ1 after exposure to EBIO (n ϭ 5; n.s.). C and D, after stimulation with 10 M forskolin, EBIO activated shrinkage in Wt cells (C), but had no effect on cftr Ϫ/Ϫ cells (D) (n ϭ 5 each). Stimulation with 10 M CCh confirmed viability of cftr Ϫ/Ϫ cells. E, EBIO did not enhance 100 nM CCh-evoked cell shrinkage (n ϭ 5). Agonists such as VIP stimulate G␣ s -mediated activation of adenylate cyclase (AC), which results in elevation of [cAMP] i and PKA-dependent phosphorylation and activation of CFTR. CFTR can function as a secretory Cl Ϫ channel, but robust secretion requires activation of counterion permeability (K ϩ conductance). Because mouse serous cells lack cAMPactivated [Ca 2ϩ ] i signals required to activate Ca 2ϩ -stimulated K ϩ channels, fluid secretion during [cAMP] i elevation is minimal. Cholinergic agonists activate G␣ q -dependent production of inositol trisphosphate (InsP 3 ), elevating [Ca 2ϩ ] i and activating Ca 2ϩ -sensitive Cl Ϫ channels (CaCCs; during strong cholinergic stimulation) and K ϩ channels (during both low-level and strong stimulation). We previously showed that mouse serous acinar cells express the CaCC Ano1 (14). Activation of both conductances is why strong cholinergic stimulation by itself can elicit a strong fluid secretion response. Fluid secretion is sustained by Cl Ϫ uptake mediated by the Na ϩ K ϩ 2Cl Ϫ cotransporter (NKCC1). Model based on this study and Refs. 15, 20, as well as the generally accepted model of exocrine fluid secretion (12). Aquaporin localization based on Refs. 42, 43. Another important goal was to shed light on functional differences observed between mouse and human serous cells and glands.
We found that while cAMP-elevating agonists activate robust CFTR-dependent anion permeability in mouse serous cells, a lack of cAMP-activated Ca 2ϩ signaling prevents cAMPagonists from stimulating robust fluid secretion due to a lack of activation of K ϩ conductance. Activation of Ca 2ϩ -activated K ϩ channels directly by 1-EBIO stimulated robust fluid secretion from forskolin-stimulated mouse serous cells. Despite a lack of significant fluid secretion generated by cAMP agonists alone, the cAMP/PKA-dependent activation of CFTR synergistically potentiates secretion during low-level cholinergic stimulation, suggesting that CFTR is an important component of the secretory pathway during times of combined cholinergic and cAMP stimulation. Our data suggest that, while the second-messenger-dependent regulation of fluid secretion differs between mouse and human or porcine serous cells, mouse serous cells are still a useful model in which to study CFTR-dependent secretion during combined cAMP and low-level cholinergic stimulation or exposure to Ca 2ϩ -activated K ϩ channel activators.
Elevation of [cAMP] i Activates CFTR-dependent Ca 2ϩ -independent Cl Ϫ Permeability in Mouse Serous Acinar Cells-Because we observed a lack of detectible serous cell fluid secretion by [cAMP] i -elevating agonists, we utilized the SPQ NO 3 Ϫ substitution assay (11,23) to determine if [cAMP] i elevation caused an activation of CFTR. Because the SPQ assay tracks electroneutral substitution of Cl Ϫ for NO 3 Ϫ , it can measure changes in anion permeability in the absence of any cation counter-ion conductance. The SPQ assay revealed large increases in CFTRdependent Cl Ϫ permeability in response to experimental [cAMP] i elevation with forskolin as well as in response to the peptide VIP and the ␤-adrenergic agonist isoproterenol. These results demonstrate functional expression of CFTR in mouse serous cells and confirm previous immunocytochemical detection of CFTR expression (15). Additionally, these data support the evidence that VIPergic and adrenergic stimulation can impact secretion from mouse submucosal glands (9) by having a direct stimulatory effect on serous acinar cell Cl Ϫ conductance. However, it appears that these agonists by themselves cannot activate robust secretion as observed during cholinergic stimulation.
Failure  (11), and ferret tracheal 3 serous acinar cells. The mechanism(s) of the generation of these Ca 2ϩ signals and the reason for their absence in mouse serous cells is not yet known, but Ca 2ϩ appears to be a required component of the secretory response to cAMP-elevating agonists in the human, porcine, and ferret cells (11). Interestingly, the importance of Ca 2ϩ signaling to cAMP-evoked fluid secretion is reflected in the increased ratio of the maximum rates of cAMPevoked to cholinergic-evoked secretion in intact porcine (8,18,19), human (6,7,16), and ferret (36) glands compared with mouse glands (9,10).
The restoration of cAMP-activated fluid secretion by EBIO suggests that cAMP-evoked [Ca 2ϩ ] i signaling is required for activation of counterion K ϩ channels necessary for secretion. The concentration of EBIO used here had no independent effects on Cl Ϫ permeability or [Ca 2ϩ ] i in serous acinar cells, suggesting that it restored secretion by direct K ϩ channel activation. EBIO activates heterologously expressed and endogenous intermediate conductance Ca 2ϩ -activated K ϩ (IK) channels with an EC 50 of ϳ75-100 M (25,26,37,38). EBIO is a weaker activator of small conductance Ca 2ϩ -activated K ϩ channels SK1-4, with an EC 50 of ϳ500 M (39). It is thus possible that the effects of 100 M EBIO observed here occur through activation of IK channels. Nevertheless, the molecular identity of these channels is yet to be identified. Future elctrophysiological studies are required to determine whether the channels activated by EBIO are the same as those activated during CCh-evoked [Ca 2ϩ ] i elevation.
In addition, the threshold [Ca 2ϩ ] i required for activation of the basolateral membrane K ϩ conductance remains to be defined. Fura-2 fluorescence changes track changes in global [Ca 2ϩ ] i , so [Ca 2ϩ ] i in the localized vicinity of the basolateral membrane K ϩ channels is unknown. In mammalian salivary acinar cells, K ϩ conductance activation slightly preceded the observed rise of [Ca 2ϩ ] i tracked by fura-2 (40), suggesting that elements of the receptor-mobilized intracellular Ca 2ϩ were localized at or near the basolateral membrane resulting in a rapid localized increase in [Ca 2ϩ ] i that exceeded the slower rise in global [Ca 2ϩ ] i . The responses to low [CCh] in mouse (this study), pig, and human (11) cells suggest that a rise of [Ca 2ϩ ] i is sufficient to activate the secretory basolateral K ϩ conductance. Coupled with the fact that forskolin and/or VIP fail to activate human or porcine cell secretion in the absence of Ca 2ϩ signaling (e.g. in BAPTA-buffered conditions; (11)), our results strongly suggest that Ca 2ϩ is the primarily second messenger responsible for K ϩ channel activation in human, pig, and mouse serous cells. Serous acinar cells from all three species appear to lack sufficient cAMP-activated K ϩ channel permeability to support secretion. The different secretory responses among these species stem solely from a lack of cAMP-induced Ca 2ϩ signaling in mouse cells and not an intrinsic difference in the regulation of the basolateral K ϩ conductance. Future studies are necessary to provide detailed insights into the identity of the K ϩ channels and their regulation in human and mouse serous cells.
We previously showed that low-level cholinergic stimulation alone is insufficient to elevate [Ca 2ϩ ] i high enough to activate robust CaCC-mediated fluid secretion in mouse, porcine, and human serous acinar cells. Here, we show that low-level cholinergic stimulation acts synergistically with cAMP stimulation to activate secretion in mouse serous cells. In forskolin-stimulated cells, the small [Ca 2ϩ ] i elevation during subsequent stimulation with 100 nM CCh is associated with marked CFTR-de-pendent cell shrinkage. This likely occurs because, despite being unable to activate CaCC, the Ca 2ϩ can activate 1-EBIOsensitive K ϩ permeability to allow secretion through the cAMP-activated CFTR Cl Ϫ permeability. This result suggests that CFTR is an important part of the secretion pathway in serous acinar cells during combined cAMP and low-level cholinergic stimulation. Stimulation with multiple agonists at low concentrations may be more reflective of in vivo physiology, where serous cells are innervated by multiple types of neurons (41).
Mouse Serous Acinar Cells As a Model of CFTR-dependent Fluid Secretion-Our data demonstrate that, in mouse nasal serous cells, CFTR-independent secretion occurs primarily during strong cholinergic stimulation, likely because [Ca 2ϩ ] i reaches sufficient levels to activate both K ϩ channels and CaCC channels. VIPergic and adrenergic stimulation alone do not elicit robust fluid secretion from serous cells because they fail to generate cAMP-activated [Ca 2ϩ ] i signals required for counterion permeability. Additionally, we found that [cAMP] i -elevating agonists do not potentiate CCh-induced Ca 2ϩ signaling in mouse serous cells, which is in contrast to porcine and human serous cells (11). However, when combined with low-level cholinergic stimulation, [cAMP] i -elevating agonists activate robust CFTR-dependent secretion. Combined cAMP-and low-level cholinergic stimulation essentially mimics the results seen when human and porcine serous acinar cells are stimulated with [cAMP] i -elevating agonists alone (11). This suggests that, while fundamental differences in the signaling pathways underlying secretion exist between mouse and human serous cells, mouse serous cells remain a useful model in which to study CFTR-dependent fluid secretion. Additionally, we have previously proposed that therapeutics to enhance cAMP-activated Ca 2ϩ signaling could result in CaCC activation and restoration of fluid secretion in CFTR-deficient human serous cells (11). Studying the comparative physiology of mouse and human serous acinar cells may yield valuable insights into the mechanisms underlying cAMP-activated Ca 2ϩ signaling in human serous cells.