cAMP-dependent recruitment of acidic organelles for Ca²⁺ signaling in the salivary gland

Abstract

Autonomic neural activation of intracellular Ca²⁺ release in parotid acinar cells induces the secretion of the fluid and protein components of primary saliva critical for maintaining overall oral homeostasis. In the current study, we profiled the role of acidic organelles in shaping the Ca²⁺ signals of parotid acini using a variety of imaging and pharmacological approaches. Results demonstrate that zymogen granules predominate as an apically polarized population of acidic organelles that contributes to the initial Ca²⁺ release. Moreover, we provide evidence that indicates a role for the intracellular messenger NAADP in the release of Ca²⁺ from acidic organelles following elevation of cAMP. Our data are consistent with the “trigger” hypothesis where localized release of Ca²⁺ sensitizes canonical intracellular Ca²⁺ channels to enhance signals from the endoplasmic reticulum. Release from acidic stores may be important for initiating saliva secretion at low levels of stimulation and a potential therapeutic target to augment secretory activity in hypofunctioning salivary glands.

the parotid gland is one of three major paired salivary glands (along with the submandibular and sublingual) that produces, elaborates, and transports saliva in response to neural input arising from the autonomic nervous system (3, 24, 58). In contrast to the canonical “gas and brake” relationship between autonomic inputs, sympathetically driven β-adrenergic receptors (β-AR) can act concurrently with parasympathetic activation of muscarinic M₃ receptors to evoke or potentiate secretion of the protein and fluid components of saliva (2, 25, 52) via inositol trisphosphate (IP₃)-mediated Ca²⁺ release from the endoplasmic reticulum (ER). However, recent literature has demonstrated that intracellular acidic organellular compartments can also act as releasable Ca²⁺ stores. Ca²⁺ signals derived from acidic organelles have been shown to regulate a wide variety of cellular functions from fertilization in starfish oocytes (39) to glucose-induced insulin release from β-cells (42) and neurite outgrowth (10). Importantly, a large content of acidic Ca²⁺ stores situated in the granular region are participants in Ca²⁺ signaling events important for fluid and digestive enzyme release from pancreatic acini (27, 50, 66). Although one particular study using submandibular acini did not find evidence of acidic organelle Ca²⁺ release (31), differences are known to exist between the major paired salivary glands. The current study examined the acidic store content of parotid acini and interrogated whether acidic organelles contribute to agonist-evoked Ca²⁺ release.

To characterize whether acidic Ca²⁺ stores contributed to agonist-evoked Ca²⁺ signals in the parotid salivary gland, we used enzymatically dissociated cultures of acinar cells in combination with immunofluorescence, live-cell fluorescence imaging, and pharmacological approaches. Our work revealed that acidic organelles are largely localized to the apical region of the acini and are predominantly comprised of secretory granules. The data indicate that acidic Ca²⁺ stores are functionally recruited following cyclic adenosine monophosphate (cAMP) elevation and suggest that nicotinic acid adenine dinucleotide phosphate (NAADP) is the second messenger that releases Ca²⁺ from secretory granules. Due to the crucial nature of Ca²⁺ signals in regulating saliva production, this pathway may have importance for understanding or treating diseases that produce salivary gland hypofunction.

MATERIALS AND METHODS

Animal use.

Male C57BL/6J mice ∼6 mo in age were obtained from The Jackson Laboratory (Bar Harbor, ME) and euthanized using CO₂ asphyxiation and subsequently punctured through the heart. All protocols involving animals were approved by the University of Toledo Institutional Animal Care and Use Committee and conformed to Guide for the Care and Use of Laboratory Animals. Access to food and water was ad libitum.

Preparation of parotid acinar clusters.

The parotid glands were quickly removed by dissection and placed in 2 ml of an enzymatic digestion solution containing 20 ml Eagle's minimal essential medium, 0.5–0.6 mg of collagenase P, 200 μl l-glutamine (2 mM), and 200 mg bovine serum albumin (1%). Glands were cleaned of fat and connective tissue, minced for several minutes until homogenous, and then placed in a 25-ml vented flask of enzymatic digest solution. Cells were incubated for 20 min with continuous shaking and gassed with CO₂ at 37°C in a water bath. The tissue was transferred to a tube and centrifuged for 1 min at 162 relative centrifugal force (rcf) with the supernatant subsequently discarded. Cells were transferred to a freshly gassed 25-ml vented flask containing the enzymatic digest solution and triturated using a 10 ml serological pipette fitted with a 200-μl tip. Tissue underwent successive 15-min digestions with shaking and gassing followed by trituration breaks using flame-polished glass pipettes with progressively smaller ends (1.19–0.6 mm). When digested to clusters containing ∼16 acini, cells were spun down for 1 min at 76 rcf, the supernatant was discarded, and the digest was halted by addition of a rinse solution containing BSA-free basal medium. After being washed, the pellet was centrifuged at 76 rcf for 1 min, the supernatant was discarded, and the pellet was resuspended in a 1:1 solution containing the rinse solution and rinse solution with 200 μl l-glutamine (2 mM) and 400 μl Penn/Strep. Parotid acinar clusters were plated on autoclaved, ethanol-wiped glass coverslips (25CIR-1), and allowed to adhere for 30 min.

Live-cell imaging.

All Ca²⁺ measurements were made at room temperature using a Nikon Eclipse TE2000S as previously described (7). Briefly, the TE2000 with DIC optics was coupled to a Polychrome IV monochromator-based high-speed digital imaging system (TILL Photonics, Gräfelfing, Germany) ported to a fiber optic guide and epifluorescence condenser. Acinar clusters were loaded with fura-2 AM, a cell-permeable ratiometric fluorescent Ca²⁺ indicator, and illuminated with dual wavelength light (340 and 380 nm) focused onto the image plane via a DM400 dichroic mirror and Nikon SuperFluor ×40 oil-immersion objective, and fluorescence was obtained through a 525 ± 25 nm band-pass filter (Chroma Technologies, Brattleboro, VT).

Immunofluorescence.

Prepared clusters of parotid acini were adhered to autoclaved glass coverslips (25CIR-1) using Cell-Tak. The acini were fixed using 4% paraformaldehyde for 15–20 min and washed three times with PBS. Permeabilization was achieved using 0.3% Triton-X-0.1% BSA for 15 min and acini were blocked to prevent nonspecific binding with 5% nonfat dry milk and 0.1% Triton-X (washing buffer) at 1 h. Antibodies for lysosomal-associated membrane protein 1 (LAMP1), vesicular-associated membrane protein 8 (VAMP8), two-pore channel 2 (TPC2), and early endosome antigen 1 (EEA1) were used at a dilution of 1:50 in washing buffer and left overnight in a humidified chamber at 4°C with mild shaking. Following incubation with the primary, the acini were washed three times in the washing buffer and Alexa Fluor secondary antibodies were prepared in the washing buffer at a dilution of 1:500 or Alexa Fluor 555-conjugated phalloidin at 1:200. Secondary antibodies were incubated at room temperature for 1 h with mild shaking. Cells were washed twice with washing buffer and once with PBS before being attached to slides on top of the antifade agent Vectashield (Burlingame, CA). After drying, slides were sealed with nail polish and imaged at room temperature on a TCS SP5 multiphoton laser scanning confocal microscope with a ×40 oil-immersion objective (Leica Microsystems).

Time-differentiated imaging.

Time-differentiated imaging was performed as described previously (7). Briefly, brightfield images were acquired every 500 ms using a high-speed Uniblitz VS35 optical shutter (Vincent Associates, Rochester, NY) placed in the tungsten lamp illumination path during recording. Images were subtracted from each previous one and exocytotic events visualized as the resultant changes in optical density.

Texas Red dextran labeling.

Isolated parotid acini plated on coverslips were incubated with 500 μl Texas red dextran solution (1 mg/ml) for 5 min. Exocytosis was then induced by adding agonist pipetted to the Texas Red solution for 10 min at 37°C. Following the stimulation period, cells were immediately fixed with 4% paraformaldehyde for 30 min. The paraformaldehyde was exchanged three times to ensure successful removal of excess Texas Red. Coverslips were washed three times at 5 min with PBS to remove paraformaldehyde and immediately imaged using confocal microscopy.

Amylase collection and Western blotting.

Parotid acinar clusters were prepared as previously described and adhered to glass coverslips. A small volume of ∼200 to 300 μl of saline was layered over the acini and amylase release was induced by addition of agonist for 10 min on ice. Following stimulation, the saline containing released amylase was collected, centrifuged for 1 min, and then frozen. The acini on the coverslips were then lysed with 200–300 μl of a buffer containing 25 mM Tris·HCl, 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, and 0.1% SDS to determine the remaining amylase content. Samples were loaded by concentration and resolved using 12% Bis-Tris gel electrophoresis. Western blotting was performed using a primary antibody for α-amylase (1:500) overnight with mild shaking at 4°C. Following incubation with the primary antibody, the membrane was incubated for 1 h at room temperature with a donkey anti-goat horseradish peroxidase-conjugated secondary antibody (1:10,000) and the blot was imaged used an ECL detection system.

Materials.

Collagenase-P used for cell preparations was bought from Roche Applied Science (Indianapolis, IN). Isoprenaline hydrochloride, carbamoylcholine chloride (carbachol), forskolin, 3-isobutyl-1-methylxanthine, and TPCN-2 primary antibody were obtained from Sigma Aldrich (St. Louis, MO). VAMP8 was acquired from Synaptic Systems (Göttingen, Germany), α-amylase, and glycyl-l-phenylalanine-β-napthylamide from Santa Cruz Biotechnology (Santa Cruz, CA), and EEA1 from Cell Signaling (Danvers, MA). LysoTracker red, Texas Red dextran, calcein green, and all secondary antibodies were purchased from Life Technologies (Eugene, OR). TefLabs (Austin, TX) supplied fura-2 AM. Bafilomyin A1 was bought from LCLabs (Woburn, MA) and NED19 acquired from R&D Systems (Minneapolis, MN).

Data analysis and image processing:.

As controls and experimental treatments were performed on different acini clusters, unpaired t-tests were used to test for statistical significance between groups with GraphPad Prism 3 (GraphPad Software, La Jolla, CA). Values represent means ± SE, with P < 0.05 taken as statistically significant. Welch's correction was used when the variances were significantly different. Bar graphs were normalized to control values for ease of presentation. All Ca²⁺ traces and subsequent analyses were done using Igor Pro software (WaveMetrics, Lake Oswego, OR). The rate of the initial Ca²⁺ rise was determined by a linear fit to the rising phase of the Ca²⁺ response. Adjustments of brightness and contrast was achieved through ImageJ (W. S. Rasband, National Institutes of Health, Bethesda, MD), and these adjustments were applied to every pixel in the image and did not obscure, eliminate, or misrepresent any information present in the original. Immunoblot data analysis was also performed using ImageJ. The relative intensity of the bands was determined and secreted amylase is expressed as percentage of total amylase.

RESULTS

Parotid acinar cells contain abundant acidic organelles with distinct distributions.

In an initial set of experiments, the acidic organelle content of parotid acini was probed using 300 nM LysoTracker red, a vital fluorescent dye that preferentially accumulates in organelles with low pH. In addition, 1 μM calcein green was used to label the cytoplasm. After 30 min of dye loading, acini were imaged using confocal microscopy. Figure 1A shows a confocal image of an acinar cell cluster with an intense red apical signal, indicating abundant acidic organelles were present at the apicoluminal pole of the acini. Occasionally ductal fragments were also observed. The pattern of LysoTracker labeling in ductal cells was less intense and typically localized to a perinuclear region (Fig. 1A, asterisk). This expression pattern indicated that there were differences in the identity of acidic organelles in acinar vs. ductal cells. Moreover, acidic organelle expression pattern varied by exocrine gland. Submandibular acini had a diffuse pattern of labeling (Fig. 1B) while lacrimal acini had abundant and polarized acidic organelles (Fig. 1C).

Fig. 1.

Acidic organelle localization and identification. LysoTracker (red, acidic organelles) and calcein (green, cytoplasm) fluorescent dyes were used for live-cell imaging (A–E) to localize acidic organelles in parotid acini (A) and a ductal fragment (A, *), submandibular acini (B), and lacrimal acini (C) and in parotid acini following treatment with BAF (D) or GPN and cathepsin inhibitor (E). The secretory granules (F) were immunolabeled using the VAMP8 antibody (green) and the subluminal F-actin network by Alexa Fluor 555-conjugated phalloidin (red). Lysosomes in acinar cells (G) and a ductal fragment (H) were visualized using LAMP1(G and H, green) and costained with the F-actin network (red). Endosomes were labeled by EEA1 antisera (I, green) with actin visualized by fluorescent phalloidin (red). All scale bars are 10 μm.

Treatment with bafilomycin A1 (BAF) to reversibly inhibit the vacuolar H⁺-ATPase and disrupt the pH gradient of acidic organelles confirmed that the labeled organelles were acidic and validated BAF as a pharmacological tool for subsequent functional studies. In these experiments, acini were incubated in 1 μM BAF for 1 h, loaded with LysoTracker and calcein, washed with fresh saline, and imaged using confocal microscopy. As shown in Fig. 1D, there was nearly a complete abolishment of the red fluorescent labeling compared with control cells treated with vehicle only. However, the identity of these organelles was not known. To further characterize parotid acinar acidic organelles, we treated cells with the lysosomal selective disruptor glycyl-l-phenylalanine-β-napthylamide (GPN) with inclusion of cathepsin inhibitor 1 to prevent off target cleavage of GPN. GPN is a substrate preferentially cleaved by the lysosomal specific protease cathepsin C to induce osmotic rupture of lysosomes. Parotid acini were treated with GPN and inhibitor for 10 min, loaded with LysoTracker and calcein, washed with saline, and imaged by confocal microscopy. Acini treated with GPN and inhibitor still displayed a robust albeit somewhat more spatially restricted LysoTracker signal (Fig. 1E). We determined the area of the cell stained by LysoTracker as a percentage of the total area of the cell stained for both control and treated acini. On average, 49% of the cell was stained by LysoTracker in control (n = 5) while 47% of the cell was stained by LysoTracker in GPN-treated acini (n = 5). This was not significantly different (P = 0.62). The overall apical pattern was generally maintained and likely represented secretory granules.

Maintenance of the LysoTracker signal following GPN treatment demonstrates that secretory granules were the more prominent acidic store in parotid acini. Accordingly, we used immunofluorescence and selective antibodies to localize a variety of acidic organelles including secretory granules, lysosomes/late endosomes, and early endosomes to ascertain the relative cell content of different acidic organelles. The vesicular-associated membrane protein 8 (VAMP8) was used as a marker for secretory granules. Figure 1F shows that acini clusters displayed immunopositive signals for VAMP8 concentrated along apicolateral surfaces, in agreement with a previous study that tested VAMP8 in parotid gland slices (62). This pattern of staining was reminiscent of the aforementioned live-cell LysoTracker images. Lysosomal stores and late endosomes were localized using lysosomal associated membrane protein 1 (LAMP1). Parotid acini displayed a pattern of LAMP1 immunoreactivity along the apical and lateral boarders of acinar cells in addition to closely apposed to the actin network (Fig. 1G). This staining mirrored our VAMP8 distribution; however, experiments assessing colocalization could not be performed as VAMP8/LAMP1 antibodies were derived from the same host. Interestingly, a ductal fragment isolated alongside the acini displayed a different LAMP1 immunoreactive pattern (Fig. 1H). This fragment displayed a more perinuclear distribution that mirrored our ductal LysoTracker image from Fig. 1A suggesting that different acidic organelles predominate between acinar and ductal cells. Although our LAMP1 studies suggests an extensive population of lysosomes, the LysoTracker fluorescent signal was largely unaffected following selective elimination by GPN treatment (Fig. 1E). This suggested the possibility that LAMP1 was on secretory granules, consistent with a previously described relationship between granules-lysosomes (8, 66). A final set of confocal images assessed the acinar endosomal distribution using the early endosome antigen 1 (EEA1) antibody (Fig. 1I). Early endosomes had a more diffuse staining pattern, with immunoreactivity in the apicolateral regions as well as some basal localization. Together, the confocal data demonstrate there are likely distinct acidic stores in acinar cells with the secretory granules comprising the majority of acidic organelles distributed apicolaterally, lysosomes/late endosomes sharing a similar distribution as well as being in close proximity to the actin network, and a more homogenous distribution of early endosomes.

Functional assessment of acidic store Ca²⁺ signaling.

Although the predominant Ca²⁺ release reservoir in most cell types is the ER (5, 28, 33) acidic organelles such as lysosomes, secretory granules, and endosomes can also function as releasable Ca²⁺ stores (49, 51). However, it is not well elucidated how these acidic stores participate in or are mobilized for global Ca²⁺ signals. Because parotid acini contained abundant acidic organelles, we began to assess the role of these organelles in agonist-evoked Ca²⁺ signals. Cells were loaded with fura-2 AM and stimulated with carbachol (CCh), an acetylcholine agonist, at concentrations we determined experimentally to evoke a threshold (50–100 nM CCh) or maximal (2 μM CCh) Ca²⁺ release (Fig. 2). It was especially important to empirically determine the threshold for CCh-evoked responses as day-to-day variations in enzymatic digestion of the tissue likely resulted in differences to cell surface receptor content between preparations. Next, the role of acidic organelles in CCh-evoked Ca²⁺ release was assessed following treatment with 1 μM BAF. Figure 3, A and B, shows that BAF-treated cells showed no significant difference in the average peak amplitude of the Ca²⁺ response at both threshold and maximal cholinergic stimulation compared with controls. Moreover, under our conditions no significant differences were observed in the average initial rate of Ca²⁺ increase or number of oscillations between control and treated cells (Table 1).

Fig. 2.

Representative Ca²⁺ traces following stimulation with varying concentrations of CCh. Ca²⁺ signals were recorded in parotid acinar cells following stimulation with the acetylcholine agonist CCh at threshold and maximal simulating concentrations (open box is saline, and black box represents concentration of CCh). Because the enzymatically dispersed cell preparations varied in their sensitivity to agonist stimulation, the threshold concentrations used to evoke Ca²⁺ signals was empirically determined. Threshold was defined as the minimum stimulus needed to typically evoke a clear initial peak release and produce oscillations (*) whose amplitudes were approximately three times the baseline noise and lasted for 2 s. Depending on the day, a threshold CCh concentration usually ranged between 50 and 75 nM. Complex Ca²⁺ waveforms were analyzed for the number of oscillations, the peak amplitude of initial Ca²⁺ rise (in ratio units or r.u), and rate of the initial Ca²⁺ rise via the slope (ratio units per second) estimated from a linear function fitted to the Ca²⁺ trace (red).

Fig. 3.

cAMP-dependent recruitment of acidic organelle Ca²⁺ release. Open box represents saline, and black box indicates duration and concentration of CCh stimulus. Acini were stimulated at threshold and maximal concentrations of CCh in the absence (A, black trace) or presence of BAF pretreatment (A, red trace). Compared with CCh alone responses [0.24 ± 0.034 r.u., n = 17], there was no change in the peak amplitude of Ca²⁺ release following BAF treatment (0.17 ± 0.031 r.u., n = 17, P = 0.16) at a threshold level of stimulation (B). Additionally, there was no significant change in peak amplitudes of CCh alone responses (1.3 ± 0.13 r.u., n = 17) compared with BAF + CCh peak amplitudes (1.2 ± 0.14 r.u., n = 17, P = 0.55) at maximal stimulation (B). However, when FORS-treated acini were stimulated at threshold and maximal concentrations of CCh with (C, red trace) or without (C, black trace) BAF pretreatment, there was a significant diminishment (D) in the threshold peak amplitude of initial Ca²⁺ release (0.11 ± 0.016 r.u., n = 6, ***P = 0.0004) compared with FORS alone responses (0.31 ± 0.035 r.u., n = 6). This was also true for maximal stimulation (D), with FORS alone peak amplitudes (2.2 ± 0.17 r.u., n = 6) significantly reduced following BAF treatment (1.6 ± 0.096 r.u., n = 6, *P = 0.013).

Table 1.

Functional assessment of acidic organelle Ca²⁺ release

Treatment	Maximal Peak Amplitude, r.u.	Threshold Rate, r.u./s	Maximal Rate, r.u./s	Threshold Oscillations	Maximal Oscillations
CCha	N/A	0.11 ± 0.027 (n = 16)	0.80 ± 0.11 (n = 16)	4.9 ± 0.59 (n = 17)	7.0 ± 0.49 (n = 17)
BAF + CCha	N/A	0.12 ± 0.042 (n = 17)	0.69 ± 0.11 (n = 17)	4.5 ± 0.81 (n = 17)	5.8 ± 0.61 (n = 17)
FORS + CChb	N/A	0.12 ± 0.023 (n = 6)	1.5 ± 0.17 (n = 6)	4.0 ± 0.63 (n = 6)	6.3 ± 0.71 (n = 6)
FORS + BAF + CChb	N/A	0.041 ± 0.012* (n = 6)	1.2 ± 0.066 (n = 6)	5.8 ± 0.65 (n = 6)	6.8 ± 0.40 (n = 6)
ISOP/IBMX + CChc	1.5 ± 0.13 (n = 13)	0.083 ± 0.023 (n = 13)	0.76 ± 0.11 (n = 13)	5.0 ± 0.74 (n = 13)	4.5 ± 0.61 (n = 13)
SOP/IBMX + BAF + CChc	1.4 ± 0.22 (n = 10)	0.025 ± 0.0084* (n = 11)	0.64 ± 0.14 (n = 10)	3.7 ± 0.82 (n = 11)	4.2 ± 0.62 (n = 10)
ISOP/IBMX + CChd	2.03 ± 0.20 (n = 6)	0.16 ± 0.019 (n = 9)	1.1 ± 0.10 (n = 6)	4.11 ± 0.51 (n = 9)	5.3 ± 0.67 (n = 6)
ISOP/IBMX 1 h + CChd	1.57 ± 0.20 (n = 8)	0.031 ± 0.0062† (n = 10)	0.85 ± 0.21 (n = 8)	1.2 ± 0.53* (n = 10)	4.6 ± 0.65 (n = 8)
ISOP/IBMX + CChe	1.8 ± 0.24 (n = 7)	0.16 ± 0.022 (n = 9)	1.2 ± 0.22 (n = 7)	3.7 ± 0.60 (n = 9)	4.7 ± 0.92 (n = 7)
ISOP/IBMX + GPN + CChe	1.5 ± 0.17 (n = 12)	0.21 ± 0.038 (n = 12)	0.83 ± 0.14 (n = 12)	3.8 ± 0.71 (n = 12)	4.4 ± 0.57 (n = 12)

Note: threshold and maximal reference the concentration of CCh used to evoke a Ca²⁺ response.

^aBAF-treated acini did not show any differences between all parameters listed.

^bFORS + CCh vs. FORS + CCh + BAF Ca²⁺ responses were not significantly altered for the different parameters except for a significant decrease in the rate of initial Ca²⁺ rise (*P = 0.011) in BAF-treated acini.

^cISOP/IBMX + CCh vs. ISOP/IBMX + CCh + BAF Ca²⁺ responses were not altered at the maximal peak amplitude, maximal rate, or at either stimulation level in terms of oscillation number. However, the rate of initial Ca²⁺ rise in ISOP/IBMX + CCh + BAF-treated acini was significantly decreased (*P = 0.039).

^dISOP/IBMX + CCh vs. ISOP/IBMX 1HR + CCh responses were unaffected at maximal peak amplitude, rate, and oscillations. ISOP/IBMX 1 HR + CCh-treated acini had a significantly decreased rate of initial Ca²⁺ rise (†P < 0.0001) and significantly fewer oscillations (*P = 0.021) under threshold signaling conditions.

^eGPN-treated acini did not show any changes at all parameters listed.

Cylic AMP-dependent recruitment of acidic store Ca²⁺ release.

Although initial experiments demonstrated acidic organelles had no observable effect on CCh-evoked Ca²⁺ signals, the aforementioned experiments assessed only activation of the parasympathetic pathway. Given that autonomic inputs are known to synergistically regulate secretory output in the salivary gland (2, 25, 52), we hypothesized recruitment of acidic organelle Ca²⁺ release depended on coincident activation of sympathetic input. Sympathetic signaling via norepinephrine activation of β-ARs increases intracellular cAMP, a messenger shown to enhance IP₃-evoked Ca²⁺ release through cAMP-activated protein kinase A (PKA)-mediated regulation (11, 60–61). Although this enhancement largely resulted from IP₃ receptor (IP₃R) phosphorylation, it is also possible the enhanced Ca²⁺ response is in part due to amplification of local acidic organelle Ca²⁺ release. Thus we designed experiments to elevate cAMP to assess its potential to recruit acidic organelle Ca²⁺ release and modulate CCh-evoked Ca²⁺ dynamics in parotid acini. First, parotid acinar clusters were preincubated for 10 min in 10 μM forskolin (FORS), an adenylate cyclase agonist, to induce a rise in cytosolic cAMP and then stimulated with CCh at either a threshold or maximal concentrations. Following BAF treatment, the average FORS-induced enhancement of the peak Ca²⁺ amplitude evoked by threshold CCh stimulation was reduced 65% compared with control responses (Fig. 3, C and D). Furthermore, the FORS-induced enhancement was diminished by 33% at a maximal CCh stimulation (Fig. 3, C and D). Additionally, there was a significant decrease in the average rate of evoked Ca²⁺ rise following BAF treatment while the number of Ca²⁺ oscillations were not altered (Table 1).

FORS robustly and globally elevates cAMP by direct activation of adenylate cyclase, but it likely does not recapitulate physiological agonist-induced changes in cAMP. Thus isoproterenol (ISOP), a β-AR agonist, was used to elevate cAMP in a more physiologically relevant manner. In these set of experiments, parotid acinar clusters were treated with 1 μM BAF for 1 h before imaging. The acini were then treated with ISOP for 5 min to elevate cytosolic cAMP and enhance Ca²⁺ release. In addition to ISOP, 3-isobutyl-1-methylxanthine (IBMX), a phosphodiesterase inhibitor was used to block cAMP degradation. As shown in Fig. 4, A and B, the average enhancement of peak Ca²⁺ release by ISOP at threshold cholinergic stimulation was reduced by ∼55% in cells that received BAF. In addition, the average rate of Ca²⁺ rise was significantly decreased in parotid acini treated with BAF (Table 1). The peak amplitude of Ca²⁺ rise was unaffected at maximal stimulation while the number of Ca²⁺ oscillations remained unaffected at both threshold and maximal CCh concentrations (Table 1). These data are consistent with the idea that acidic organelle recruitment and Ca²⁺ release required agonist-evoked elevation of cAMP.

Fig. 4.

ISOP evoked elevations in cAMP. Open box represents saline, and black box indicates duration and concentration of CCh stimulus. Acini were stimulated at a threshold concentration of CCh following elevation of cAMP via ISOP/IBMX in the absence (A, black trace) or presence (A, red trace) of BAF pretreatment. Compared with ISOP/IBMX alone responses (0.27 ± 0.048 r.u., n = 13), there was a significant reduction in the peak amplitude of Ca²⁺ release following BAF treatment (0.13 ± 0.029 r.u., n = 11, *P = 0.020) at a threshold level of stimulation (B). Responses to maximal stimulation were not significantly different following BAF treatment.

Secretory granules are the relevant acidic organelle Ca²⁺ release site.

While use of BAF revealed a role for acidic Ca²⁺ stores in Ca²⁺ signaling, BAF treatment was relatively unselective and ubiquitously dissipated all the acidic stores. Thus additional functional studies were performed using more selective disruptors of acidic organelles to further probe the contribution of individual acidic compartments to the observable Ca²⁺ signal. To address whether secretory granules served as a predominant site for acidic store Ca²⁺ release, we treated acini with prolonged ISOP stimulation (1 h) to deplete granular stores. This degranulation was confirmed using live-cell calcein/LysoTracker staining (Fig. 5A). Ca²⁺ signals were assessed once again at threshold and maximal levels of cholinergic stimulation following extended ISOP treatment and compared with control acini that underwent the standard 5–10 min ISOP treatment. ISOP treatment for 1 h followed by stimulation at a threshold concentration of CCh produced a 66% diminishment in the peak amplitude of Ca²⁺ release (Fig. 5, B and C). Additionally, the rate of Ca²⁺ rise was significantly decreased following secretory granule depletion and the number of Ca²⁺ oscillations were also diminished at threshold cholinergic stimulation (Table 1). Secretory granule depletion had no significant effects on Ca²⁺ signaling dynamics when acini were stimulated at a maximal CCh concentration (Table 1).

Fig. 5.

Secretory granules appeared to be the dominant acidic organelle Ca²⁺ release site. Confocal images show acinar cells (A) labeled with LysoTracker (red, acidic organelles) and calcein (green, cytoplasm). Compared with a control treated cell, prolonged stimulation with ISOP for 1 h largely depleted the secretory granule population indicated by a nearly complete abolishment of the LysoTracker signal. Scale bar is 7.5 μm for control and 10 μm for ISOP 1 h. Following depletion of secretory granules by ISOP treatment for 1 h, acini were stimulated with a threshold concentration of CCh (B, red trace) and compared with acini treated with ISOP/IBMX for 10 min (B, black trace). Open box represents saline, and black box indicates duration and concentration of CCh stimulus. Granule depleted acini had diminished peak amplitudes (0.15 ± 0.027 r.u., n = 10) compared with controls (0.44 ± 0.034 r.u., n = 9, ***P = <0.0001) at threshold stimulation (C). Responses to maximal stimulation were unaffected. GPN-treated acini were stimulated with a threshold concentration of CCh (D, green trace) and compared with acini treated with ISOP/IBMX for 10 min (D, black trace). Lysosome-depleted acini at threshold stimulation (E) had peak amplitudes (0.44 ± 0.048 r.u., n = 12) that were not significantly different from control acini (0.45 ± 0.038 r.u., n = 9, P = 0.72). Responses to maximal stimulation were unaffected.

Imaging experiments indicated that lysosomes did not contribute significantly to the overall LysoTracker (or acidic organelle) signal. However, it was still important to assess the functional consequences of depleting lysosomal stores. In these experiments we used GPN in the presence of cathepsin inhibitor 1 to osmotically rupture lysosomes. Fura-2 AM loaded acini were then treated with ISOP/IBMX to elevate cAMP and stimulated with threshold and maximal concentrations of CCh. Compared with controls, GPN-treated acini did not exhibit significant changes in the peak amplitude of Ca²⁺ rise under threshold CCh stimulation (Fig. 5, D and E). This remained true for all examined analytical parameters at either CCh concentration (Table 1). Interestingly, there was a subtle but statistically significant increase in the resting 340/380 ratio following GPN treatment (1.4 to 1.6, P = 0.0019). However, whether this increase affected CCh-evoked Ca²⁺ responses was not determined. Therefore, the significant reduction in the Ca²⁺ signal following secretory granule depletion, coupled with minimal changes in Ca²⁺ signals observed in GPN-treated cells, identified secretory granules as a dominant site for acidic organelle Ca²⁺ release within parotid acini.

NAADP as a potential mediator of acidic store Ca²⁺ release.

Having demonstrated that acidic organelles participated in the enhancement of CCh-evoked Ca²⁺ signals at threshold levels of stimulation following cAMP elevation, we tested the role of a previously identified mediator of acidic store Ca²⁺ release-the potent small molecule NAADP. NAADP was first discovered in sea urchin eggs to target an IP₃, cyclic ADP ribose insensitive Ca²⁺ store (38), eventually shown to be an acidic lysosomal-related organelle (19). Moreover, the putative enzyme that converts NADP to NAADP has been shown to be regulated by cAMP in other cell types (12, 54, 65). To test the hypothesis that NAADP contributed to Ca²⁺ release from acidic stores, parotid acini were incubated for 30 min with 10 μM NED19, a NAADP receptor antagonist, and stimulated with either a threshold or maximal concentration of CCh. Treatment with NED19 had no effect on Ca²⁺ dynamics in acini stimulated with CCh alone at either concentrations of cholinergic agonist compared with controls (Fig. 6, A and C, and Table 2). However, when ISOP was used to elevate cAMP, NED19 reduced the CCh-evoked peak Ca²⁺ release by 40% at threshold stimulation compared with cells treated with ISOP alone (Fig. 6, B and C). Similar to BAF pretreatment, the rate in which Ca²⁺ reached its peak amplitude was significantly slower in NED19/ISOP-treated acini (Table 2) at threshold CCh-evoked signals but without affect at maximal stimulation. The number of oscillations was unaffected at threshold and maximal CCh concentrations (Table 2). The significant NED19-induced decrease in the threshold Ca²⁺ response mimicked BAF treatment and was consistent with other studies that identified NAADP as a second messenger releasing Ca²⁺ from acidic organelles.

Fig. 6.

NED19-dependent inhibition of Ca²⁺ release. Open box represents saline, and black box indicates duration and concentration of CCh stimulus. Acini were stimulated at a threshold concentration of CCh with (A, red trace) or without (A, black trace) NED19 pretreatment. There was no change in the peak amplitude of Ca²⁺ release following NED19 treatment (0.13 ± 0.038 r.u., n = 10) when compared with control (0.17 ± 0.032 r.u., n = 14, P = 0.44) at a threshold level of stimulation (C). In contrast, following cAMP elevation in the absence (B, black trace) or presence (B, red trace) of NED19 pretreatment, there was a significant reduction in the peak amplitude (ISOP/IBMX, 0.14 ± 0.017 r.u., n = 13 vs. ISOP/IBMX + NED19, 0.088 ± 0.016, n = 13, *P = 0.026) of threshold responses (C).

Table 2.

NED19-dependent inhibition of Ca²⁺ release

Treatment	Maximal Peak Amplitude, r.u.	Threshold Rate, r.u./s	Maximal Rate, r.u./s	Threshold Oscillations	Maximal Oscillations
CCh*	1.1 ± 0.081 (n = 14)	0.043 ± 0.011 (n = 14)	0.52 ± 0.058 (n = 14)	3.6 ± 0.68 (n = 14)	4.9 ± 0.73 (n = 14)
CCh + NED19*	0.95 ± 0.12 (n = 10)	0.026 ± 0.010 (n = 10)	0.49 ± 0.079 (n = 10)	3.2 ± 0.63 (n = 10)	4.7 ± 0.79 (n = 10)
ISOP/IBMX + CCh†	0.97 ± 0.087 (n = 13)	0.044 ± 0.012 (n = 13)	0.62 ± 0.099 (n = 13)	3.2 ± 0.78 (n = 13)	4.9 ± 1.0 (n = 13)
ISOP/IBMX + NED19 + CCh†	0.90 ± 0.069 (n = 13)	0.016 ± 0.0049‡ (n = 13)	0.41 ± 0.056 (n = 12	2.1 ± 0.60 (n = 13)	3.1 ± 0.66 (n = 13)

Note: threshold and maximal reference the concentration of CCh used to evoke a Ca²⁺ response.

^*NED19-treated acini had no significant differences between the listed parameters.

^†The rate of initial Ca²⁺ rise at threshold stimulation was significantly decreased in ISOP/IBMX + CCh + NED19-treated acini (‡P = 0.044) with no other observable differences.

Presence of acidic store Ca²⁺ release machinery.

Two-pore channels (TPCs) have been implicated as the lysosomal Ca²⁺ channel responsible for acidic store Ca²⁺ release (9, 13, 73). There are three isoforms of TPCs (TPC1-3) and RNA message for both TPC1 and TPC2 has been found in salivary tissue, although only the TPC2 protein has shown to be expressed. Using immunoflourescent methods and confocal microscopy we determined the subcellular distribution of TPC2. The acini clusters displayed punctate immunoreactivity for TPC2 (Fig. 7A) prominently along the lateral and apical membranes with very minimal staining basally. We next investigated redistribution of TPC2 following inducement of secretory granule exocytosis. To achieve this, cells were treated with 4 μM ISOP for 30 or 60 min and then stained for TPC2. Thirty minutes of ISOP treatment resulted in a gradual increase of TPC2 immunopositive staining (Fig. 7B); however, treatment for 60 min produced a robust increase in the signal of TPC2 along the apical and lateral membranes (Fig. 7C). Our results identify that TPC2 redistributes following prolonged ISOP treatment, suggesting TPC2 might be localized on secretory granules that are known to redistribute following exocytosis.

Fig. 7.

The presence of TPC2. TPC2 (green, A) was subcellularly distributed along the apical and lateral boarders of acinar cells, a pattern reminiscent of the LysoTracker staining and VAMP8 distribution. Attempts were made to determine whether TPC2 was redistributed following exocytosis. Cells were pretreated with ISOP for 30 min (B) or 1 h (C) before labeling. There was some accumulation of TPC2 along the apical and lateral boarders after ISOP treatment for 30 min (B, *) and robust accumulation following ISOP treatment for 1 h (C, *). All scale bars are 5 μm.

Downstream effects of disrupted acidic organelle Ca²⁺ release.

Our results show acidic organelle Ca²⁺ release contributes to an agonist-evoked global Ca²⁺ signal in parotid acinar cells. These Ca²⁺ signals are known to drive a variety of downstream processes such as control of the exocytotic release of salivary proteins and activation of Ca²⁺-regulated ion channels coordinating fluid transport. To explore whether acidic organelle Ca²⁺ release was important for regulating salivary protein release, experiments were performed to assess whether CCh + ISOP-induced exocytosis from acini was altered following pretreatment with 1 μM BAF for 1 h. To assess exocytotic function, we used a variety of techniques that included 1) time-differentiated imaging, 2) Texas Red dextran labeling, and 3) Western blotting for secreted amylase. To quantify data from time-differentiated imaging, the number of exocytotic events were normalized to the number of acini present during recording. During a threshold level of CCh stimulation (30 s) following pretreatment with ISOP/IBMX, acini displayed 1.52 events/cell, which was slightly albeit unsignificantly reduced in the presence of BAF (1.10 events/cell, Fig. 8A). Due to potential difficulties discerning changes in optical density, additional experiments measured exocytosis by uptake of a dextran conjugated to Texas Red. This dextran was used as a fluid phase probe that was likely incorporated via fusion pores during exocytosis (and endocytosis). To identify sites of exocytosis, control or treated cells were fixed and the exocytotic sites revealed by Texas Red fluorescence using confocal microscopy (Fig. 8B). Images were quantified and the percentage of Texas Red signal to total area of the cell was determined. At thresholds level of CCh stimulation in the presence of ISOP, acini had ∼8.5% of total area stained with Texas Red, which was not significantly different following BAF treatment (11.1%, Fig. 8C). A final set of experiments quantified amylase release with or without BAF pretreatment using Western blotting (Fig. 8D). Acini were stimulated with a high concentration of CCh in the presence of ISOP to promote maximal amylase release. CCh+ISOP stimulated acini secreted ∼51.5% of the total amylase content. Consistent with our other assessments of exocytotic activity, CCh + ISOP stimulated acini pretreated with BAF did not appear to have a significantly altered secretion of the total amylase content at 45.2%. Interestingly, although there was a significant effect on a global Ca²⁺ signal following disruption of acidic organelles it appeared this change did not alter the exocytotic profile. It is possible that acidic organelle Ca²⁺ signaling may selectively modulate salt secretion in parotid acini.

Fig. 8.

Acidic organelle Ca²⁺ release does not affect exocytosis. A: using time-differentiated imaging, disruption of acidic organelle Ca²⁺ signaling with BAF did not significantly diminish exocytosis at a threshold level of CCh stimulation following cAMP elevation with ISOP/IBMX (n = 9 for ISOP/IBMX + CCh, n = 8 for ISOP/IBMX + CCh + BAF, P = 0.3905). B and C: whereas there was an increase in the percent of Texas Red dextran labeling of acini when the cells were stimulated with ISOP and CCh indicating exocytotic activity (B, scale bars are 10 μm), BAF treatment did not diminish the agonist-evoked signal (C, n = 5 for all groups, P = 0.3762). D: consistent with imaging experiments, Western blotting for secreted amylase revealed that the percentage of total amylase released increased significantly following ISOP and CCh costimulation for both control and BAF-treated acini (n = 4 for both groups, P = 0.6943). Inset (D) is a representative Western blot for amylase. Control refers to ISOP + CCh-treated cells, and BAF-treated represents acini stimulated with ISOP + CCh following preincubation with BAF. S, secreted amylase; L, lysed amylase fraction.

DISCUSSION

Generation of Ca²⁺ signals is a complex and highly regulated process in terms of time, space, and converging messengers. To date, considerable effort has focused on intracellular Ca²⁺ signals that control the release of Ca²⁺ from the ER and ER resident Ca²⁺ channels. However, recent studies have begun to define the role of acidic organelles as a releasable Ca²⁺ store, thus adding an additional layer of regulation to an already inherently complex system.

In the current study, the identification, subcellular distribution, and contribution of acidic organelles to agonist-evoked Ca²⁺ dynamics was investigated in parotid salivary gland acinar cells. Acidic organelles were prominently localized to the apical membrane and subsequent immunofluorescent labeling showed these stores included secretory granules, lysosomes, and endosomes. A surprising and important finding of this study was a coincident signal in the form of cAMP was necessary to mobilize release of Ca²⁺ from acidic organelles and modulate Ca²⁺ signals evoked by threshold cholinergic stimulation. Threshold signaling represents the modest stimuli likely predominating under most physiological situations and may indicate acidic stores act as an important sensitizing factor for the initiation of saliva secretion. In contrast, we hypothesize that the contribution by acidic organelles during maximal CCh stimulation is overwhelmed by large-scale ER release and store operated Ca²⁺ entry important for sustained fluid production (1, 44–46, 53).

The observation of cAMP-dependent recruitment of acidic organelles in the parotid salivary gland aligns with the idea that specific agonists may produce specific signatures of Ca²⁺ signals (6, 14–16, 27, 29, 66–67). Previous work in pancreatic acini demonstrated that Ca²⁺ signaling patterns derived from the circulating hormone cholecystokinin were induced by NAADP (15–16, 67). On the other hand, acetylcholine-evoked Ca²⁺ release from a nonacidic pool (27, 66) and did not significantly elevate intracellular NAADP (67). We propose in parotid acinar cells that β-AR activation elevated NAADP and induced Ca²⁺ release from an acidic pool whereas CCh alone elicited signals were unaffected by BAF or NED19 treatment as IP₃ production primarily released Ca²⁺ from a nonacidic, thapsigargin-sensitive store. Thus parotid acini, like pancreatic acini, can use agonist specific intracellular messengers to coordinate signal specificity or intensity.

Although the current study suggested acidic organelle Ca²⁺ release is initialized by a rise in NAADP following elevation of cAMP, the intermediaries of this pathway are unknown. A likely candidate for activation of the putative enzyme for NAADP production is PKA, shown previously to regulate ADP-ribosyl cyclase activity, a NAADP synthesizing enzyme (35, 54, 56). However, experiments using inhibitors of PKA, such as PKI, would presumably not only block acidic store Ca²⁺ release but also block PKA-dependent phosphorylation of IP₃Rs. The type II IP₃R subtype is the most abundant IP₃R in parotid acinar cells (71) and in acini have been shown to be phosphorylated by PKA (11), which results in enhanced Ca²⁺ release. Thus we were unable using current approaches to distinguish the effects of PKI on acidic store Ca²⁺ release vs. PKI's dampening effect on IP₃R phosphorylation. Another possibility is that cAMP can also activate proteins known as exchange proteins directly activated by cAMP (EPAC) that are known to regulate Rap1 (a guanine nucleotide exchange factor). Indeed, EPAC has been shown to sensitize r yanodine receptors to accelerate Ca²⁺ responses in pancreatic acini (55) and recent work by our lab showed EPAC activation can enhance Ca²⁺ signals evoked by purinergic receptor activation via a CICR mechanism (6).

Although localization of TPC2 in parotid acini to a region similar to acidic organelle distribution would suggest these proteins are candidates for mediating acidic organelle Ca²⁺ release, the literature is equivocal when it comes to whether TPCs are the primary site of action by NAADP. Work by Xu and colleagues (63) also argues that TPCs are in fact sodium, not Ca²⁺, conducting channels. Moreover, several studies demonstrated NAADP can mobilize Ca²⁺ through ryanodine receptors (26, 32, 47) or through closely related mucolipin channels of the acidic endolysosomal system (4, 69–70). Interestingly, demonstrating NAADP's binding site on TPCs remains elusive. An alternative hypothesis is that NAADP, through a binding partner (40), can regulate the variety of ion channels shown to be important for NAADP-induced Ca²⁺ signaling (30). Nevertheless, there is a general consensus that TPCs are important for NAADP-mediated Ca²⁺ release from an acidic pool.

Functionally, secretory granules appeared to be the predominant site of acidic organelle Ca²⁺ release as acini depleted of secretory granules produced diminished Ca²⁺ signals whereas osmotic rupturing of lysosomes had no discernable effect. However, the contribution of endosomes could not be directly assessed using current pharmacological approaches. Other groups have addressed the contribution of zymogen granules to a global Ca²⁺ signal by inducing depletion via intraperitoneal injection of ISOP (48) or parotid ductal ligation (41). Their results indicated there were no observable changes in a CCh-evoked Ca²⁺ response and suggested secretory granules likely were not key players in generating Ca²⁺ signals. However, the former study used CCh concentrations that produced robust rather than threshold Ca²⁺ signals. In contrast, our results support a role for acidic organelle Ca²⁺ release only at threshold stimulation and this signal is less important or can be bypassed at stronger stimulation. In the work by Liu et al. (41) experiments probed threshold CCh signaling, however, this group performed experiments in the absence of ISOP or other cAMP elevators. Taken together, we hypothesize secretory granules function as acidic Ca²⁺ stores that are mobilized for Ca²⁺ release following agonist-evoked rises in cAMP. We believe localized Ca²⁺ release can sensitize proximally located canonical channels, IP₃Rs and ryanodine, allowing for enhanced release from the ER upon activation by IP₃, consistent with the “trigger” model for acidic organelle Ca²⁺ release (15, 27).

Experiments revealed that although release of Ca²⁺ from acidic secretory granules enhanced the release from ER Ca²⁺ stores, the abolishment of this “trigger” using BAF did not result in any downstream changes to exocytosis. This finding was unexpected as studies in salivary gland acini demonstrated changes in pH to granules using weak bases (our experiments used BAF) affected the retention of proteins in mature secretory granules (59) and induced more basolateral secretion (34). Moreover, studies in T cells have shown that release of Ca²⁺ from acidic cytolytic granules through NAADP acts in an autocrine fashion to regulate release of the granules themselves (21–22). One caveat is that the contributions of acidic organelle Ca²⁺ release may have been overwhelmed by cAMP-dependent exocytosis which promotes the bulk of exocytosis in parotid acini.

The findings presented here may have clinical relevance for those suffering from Sjogren's Sydrome, an autoimmune disease characterized by dry mouth and dry eye. Early stage effects on the parotid include decreased cholinergic sensitivity (23, 36), a subsequent decrease of cholinergic-induced IP₃ signaling, reduced Ca²⁺ release, and ultimately diminished saliva production i.e., chronic dry mouth. Interestingly, a recent publication by Teos et al. (57) also demonstrated IP₃R decreases in Sjogren's patients that reduced CCh-evoked Ca²⁺ mobilization. Although the acidic organelle population in the minor salivary glands used in the Teos et al. study was unknown, our work suggests that Ca²⁺ release from acidic organelles could be selectively used as a pharmacological target to sensitize any residual IP₃Rs to augment Ca²⁺ release to in turn increase fluid secretion, and reduce symptoms of dry mouth. The parotid is not the only exocrine gland targeted by Sjogren's. Initial results to test the acidic organelle population in submandibular and lacrimal glands showed differential staining patterns. Abundant and polarized acidic organelles were seen in lacrimal tissues while submandibular acini had a more diffuse pattern of labeling. This diffuse pattern of labeling may explain results from Harmer et al. (31), who demonstrated NAADP did not evoke a K⁺ or Cl⁻ current response indicative of Ca²⁺ release, presumably because of the sparse acidic organelle population in submandibular acini. Future work is ongoing to characterize the functional recruitment of the abundant acidic organelles in lacrimal acini. Acidic organelles may have a key role in regulating fluid secretion and offer therapeutic targets to treat dry eye in addition to dry mouth.

GRANTS

This work was funded by National Institutes of Health National Institute of Dental and Craniofacial Research Grant DE-023418.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

J.F.I. and D.R.G. conception and design of research; J.F.I., S.B., S.K., A.W., P.G., and A.K.I. performed experiments; J.F.I., S.B., S.K., A.W., P.G., and A.K.I. analyzed data; J.F.I., S.B., S.K., A.W., P.G., and D.R.G. interpreted results of experiments; J.F.I. prepared figures; J.F.I. drafted manuscript; J.F.I., S.B., and D.R.G. edited and revised manuscript; J.F.I. and D.R.G. approved final version of manuscript.