Effect of P2X₇ receptor knockout on exocrine secretion of pancreas, salivary glands and lacrimal glands

Materials

All standard chemicals including collagenase, hormones and agonists were obtained from Sigma (Copenhagen, Denmark). Tissue culture media and phosphate-buffered saline (PBS) were from Gibco/Invitrogen (Denmark). Luciferase and luciferin were from Roche Diagnostics (Germany). Mebumal was from Nycomed (Roskilde, Denmark).

Ethical approval

The permission for animal experiments, including the below described protocols, and breeding of transgenic mice was obtained from the Danish Animal Experiment Inspectorate (Dyreforsøgstilsygnet). The animal experiments also comply with the guidelines adopted by The Journal of Physiology (Drummond, 2009). In order to minimize the number of animals, in vivo and in vitro functions of the three glands were studied simultaneously (see below).

In vivo collection of exocrine secretion

P2X₇^−/− mice on C57BL/6 background were originally obtained from Pfizer (Groton, CT, USA) and wild-type mice of the same strain were obtained from Taconic (Ejby, Denmark) and used for breeding. Mice were housed in standard animal house and had access to chow and water ad libidum. We used age- and sex-matched wild-type and receptor knockout animals. They were 20–40 weeks old and weighed 25–40 g. The total number of animals used for this part of the study was 19. For in vivo experiments mice were anaesthetized with Mebumal (pentobarbital, 5 mg per 100 g body weight i.p.) and anaesthesia was maintained during the experiments by additional injections of Mebumal. Mice were placed on a heated surgical table, the rectal temperature was monitored and animal was maintained at 38°C. A tracheal cannula was inserted to avoid aspiration of mucus and saliva. The abdomen was opened by a midline incision and the pylorus and the proximal end of the bile duct were ligated. The common pancreatic bile duct was cannulated with polyethylene tube and pancreatic juice was collected into pre-weighed vials at timed intervals before and after stimulation. Mixed saliva was collected from the mouth at timed intervals. Tears were collected into a polyethylene tube placed at the corner of the eyes and secretion rate was calculated from the volume of fluid collected. Secretion in all glands was evoked by pilocarpine (1 mg (100 g)⁻¹, i.p.). At the end of the experiment, animals were killed by an overdose of Mebumal and exocrine glands (pancreas, submandibular, parotid, sublingual and lacrimal glands) and other vital organs were removed, weighed and examined. Secretion rates were corrected per gram of gland weight. Corrections for the body weight gave similar results.

Preparation of cell suspensions

For in vitro measurements of ATP release and intracellular Ca²⁺ activities, cell suspensions of three major glands were prepare simultaneously from a single mouse. Mice were killed with cervical dislocation, and pancreas, parotid and exorbital lacrimal glands were removed, placed into cold PBS supplemented with 0.25 mg ml⁻¹ trypsin inhibitor and cut into small pieces (<1 mm³). In the lab, solutions were replaced with incubation medium based on Ham's F12/Dulbecco's modified Eagle's medium (DMEM)-1000 mix: 5 ml for pancreas, 3 ml for parotid glands and 2 ml for lacrimal glands. In addition, media contained 1.33 mg ml⁻¹ collagenase (type V) (pancreas) or 2.66 mg ml⁻¹ (parotid and lacrimal); 0.1 mg ml⁻¹ hexokinase or 3 U ml⁻¹ apyrase, 0.165 mg ml⁻¹ trypsin inhibitor; 0.16 mg ml⁻¹ hyaluronidase (lacrimal and parotid preparations) and 1 mm EGTA. Tissues were incubated in rotating bath at 37°C and 5% CO₂ in O₂. At 10 min intervals tissues or cells were dispersed by pipetting and incubation was terminated at 20 min (pancreas) or 30 min (parotid and lacrimal glands) by addition of a cold medium. Cell suspensions were washed and passed through a nylon filter according to procedures described earlier (Hede et al. 1999). With this method it was possible to prepare intact acini/small cell clusters and small ducts. For further experiments cells were gently washed in physiological HCO₃⁻-free buffer (–BIC) of the following composition (in mmol l⁻¹): Na⁺ 145, K⁺ 3.6, Ca²⁺ 1.5, Mg²⁺ 1, Cl⁻ 145, phosphate 2.0, glucose 5 and Hepes 10.

ATP measurements

Cells suspended in –BIC solution were pipetted in 50 μl aliquots into 96-well microtitre plates followed by 50 μl of luciferin–luciferase mix from assay kit HSII (Roche Diagnostics, Manheim, Germany) or FL-AA (Sigma), which were dissolved in –BIC. Cells were allowed to rest for 45–60 min. In order to prevent nucleotide hydrolysis (Yegutkin et al. 2006), the following nucleotidase inhibitors were also added to each well: 0.3 mmβ,γ-methylene-ATP and 2 mm levamizol. Luminescence was monitored directly in cell wells in FLUOStar optima (BMG Labtech, Offenburg, Germany) at 25°C. Luminescence was monitored in 1 s intervals prior to and after injection of 5 μl volumes of –BIC (control) and then an agonist. ATP standards were treated as samples and standard curves were constructed for each experiment. ATP release in response to given agonist was calculated from mean peak responses (10 s) corrected for the control –BIC response due to the mechanical disturbance by the pump injection. ATP release, monitored in arbitrary luminescence units, was recalculated as ATP concentrations. Cells were lysed (according to kit protocols) and assuming that all ATP released originates from intracellular ATP of about 3 mm, one can calculate the number of viable cells. Cell numbers obtained by this method were verified by haematocytometer count in duplicate samples. Concentrations of ATP released from cells following agonist stimulation were corrected for 1 million cells for each sample. The following agonists were used: carbachol (50 μm), cholecystokinin octapeptide (CCK-8; 7 × 10⁻¹⁰m), neurotensin (1.25 × 10⁻⁷m) and pilocarpine (0.4 mm).

Fura-2 measurements

Cells suspended in –BIC were also used to measure changes in intracellular Ca²⁺ activities using Fura-2. It is well accepted that in exocrine gland cells, agonists increase intracellular Ca²⁺ activity (due to release of Ca²⁺ from intracellular stores and Ca²⁺ influx), and these lead to secretory events, such as stimulation of Ca²⁺ sensitive Cl⁻ and K⁺ channels, and eventually ion and fluid secretion. For simplicity, in the following text we shall refer to changes in intracellular Ca²⁺ activity as Ca²⁺ signals. Fura-2 measurements were made in the microtitre plate reader FLUOStar Optima. In order to minimize movement of cells during pump injections, cells were held on cut-out small filters (Poretics, Sterlitech, Kent, WA, USA) attached with medical adhesive (Hollister, Dansac & Hollister, Fredensborg, Denmark) to the bottom of wells. Cells were loaded with 5 μm Fura-2/AM (Invitrogen) for 30–45 min, then washed and re-suspended in 200 μl of –BIC solution in wells. After further equilibration, Fura-2 signal emission of cells in 3 mm × 3 mm matrixes was measured at 510–520 nm after excitation at 340 and 380 nm. Cells were stimulated by injection of 5 μl of agonist. The following agonists were used: carbachol (50 μm) and ATP or ATPγS (0.1 mm). The peak change in Fura ratio with a given agonist was taken as a measure of intracellular Ca²⁺ signals.

Microscopy

Cell suspensions of exocrine glands were examined with a confocal laser scanning microscope (Leica SP CLSM equipped with an Ar–Kr laser) with 20× 1.7 NA HC PL APO and 63× 1.2 NA PL APO objectives. In order to estimate ATP stores, cells were incubated with 1–5 μm quinacrine dihydrochloride for 5–15 min and fluorescence was detected at 490–540 nm with 476 nm excitation (Sørensen & Novak, 2001). Exocrine glands were also fixed in 4% parafomaldedyhe and stained with haematoxylin/eosin and general morphological structure was examined by light microscopy.

Data presentation

Data are presented as original recordings and summaries showing the mean values ±s.e.m. with n indicating a number of independent experiments. For analysis of responses ANOVA or Student's t test was used. Data were analysed in Origin (OriginLab Corp., Northampton, MA, USA).

Pancreatic, salivary and lacrimal gland secretions

P2X₇ receptors have various effects on exocrine glands at a cellular and subcellular level. In order to test whether these receptors also affect exocrine secretion, we set up in vivo experiments on P2X₇^+/+ and P2X₇^−/− mice that were stimulated with pilocarpine, a muscarinic receptor agonist that is a good secretagogue of epithelial secretion. Simultaneously, we collected pancreatic secretion, saliva and tears prior to and after stimulation and the results are shown in Figs 1–3. Secretory rates were corrected for the gland weight, as at the end of the experiments pancreas, major salivary glands (submandibular, parotid and sublingual) and exorbital lacrimal glands were dissected and weighed. Similar results as depicted in Figs 1–3 were obtained when secretory rates were corrected for animal weight.

Figure 1 shows secretion rates (per gram of gland) of pancreatic juice of P2X₇^+/+ and P2X₇^−/− mice as a function of time. Non-stimulated secretion was relatively low, and pilocarpine evoked relatively high secretion rates in both wild-type and receptor knockout animals. In a relatively large number of animals of both sexes, it became apparent that after the initial secretory response (10–20 min), pancreas of P2X₇^−/− animals appeared to secrete at about 30% lower rate (Fig. 1A). In order to determine whether there were any sex differences, we analysed the data as shown in Fig. 1B and C. Clearly, the female pancreas of all animals secreted at lower rates than the male pancreas, and there seemed to be no differences between the wild-type and knockout mice. Integrated 60 min secretion was 242 ± 51 μl g⁻¹ h⁻¹ in wild-type and 234 ± 82 μl g⁻¹ h⁻¹ in knockout female mice (n = 6 and 7). Wild-type males secreted significantly more pancreatic juice (per gland or body weight), and knockout of the receptor appeared to decrease secretion. Integrated 60 min secretion was 438 ± 69 μl g⁻¹ h⁻¹ in wild-type and 339 ± 72 μl g⁻¹ h⁻¹ in knockout male mice (n = 7 and 8). This effect was more pronounced after prolonged stimulation, i.e. mean secretion at 20–60 min was 7.46 ± 0.21 μl g⁻¹ min⁻¹ in wild-type and 4.82 ± 0.18 μl g⁻¹ min⁻¹ in P2X₇^−/− male mice (n = 7, 8), which was 35% lower (P = 8.8 × 10⁻⁵).

Compared to pancreas, 60 min salivary secretion was much higher and amounted to 1341 ± 154 μl g⁻¹ h⁻¹ in all wild-type mice and 931 ± 91 μl g⁻¹ h⁻¹ in P2X₇^−/− mice (n = 17 and 19; P = 0.024). Figure 2 shows salivary secretion rates as a function of time; the typical biphasic response may be due to different sets of ion transporters contributing to net secretion with prolonged stimulation (Novak & Young, 1986). Again in order to detect possible sex differences, data are plotted for male and female mice in Fig. 2B and C. In the peak phase of secretion (arbitrarily defined integrated 0–30 min of secretion), P2X₇^+/+ male mice secreted significantly higher than P2X₇^−/− mice, i.e. 990 ± 110 vs. 626 ± 74 μl g⁻¹ 0.5 h⁻¹ (P = 0.0237). In the plateau phase (30–60 min), the total secretion was 267 ± 63 μl g⁻¹ 0.5 h⁻¹vs. 114 ± 39 μl g⁻¹ 0.5 h⁻¹ in the same animals (n = 9 and 8). In female animals, the initial integrated peak secretion (0–30 min) was similar, i.e. 1114 ± 205 μl g⁻¹ 0.5 h⁻¹ in wild-type and 960 ± 110 μl g⁻¹ 0.5 h⁻¹ in receptor knockout animals. In the same animals, the plateau integrated secretion (30–60 min) was 323 ± 85 μl g⁻¹ 0.5 h⁻¹ in P2X₇^+/+ mice and significantly lower 109 ± 31 μl g⁻¹ 0.5 h⁻¹ in P2X₇^−/− animals (n = 8 and 10; P = 0.021). Taken together, the secretory patterns were similar in salivary glands of both male and female mice (Fig. 2B and C). However, in the peak phase of secretion, male glands appeared to be more affected by the receptor knockout, while in the plateau phase of secretion (30–60 min), all knockout animals had secretion of about half that observed in wild-type animals, i.e. mean secretion at 30–60 min was 9.94 ± 1.68 μl g⁻¹ min⁻¹ in wild-type and 3.81 ± 0.79 μl g⁻¹ min⁻¹ in P2X₇^−/− mice (n = 17 and 18), which was 62% lower (P = 0.019).

Tears are produced by several glands/epithelia, and the most significant contribution is from the lacrimal glands. Figure 3A shows that pilocarpine stimulation increased tear production 3.8-fold in wild-type animals, i.e. from 2.07 ± 0.88 to 7.95 ± 1.35 μl g⁻¹ min⁻¹ (n = 16; P = 0.002), and 4.3-fold in knockout animals, i.e. 2.49 ± 0.81 to 10.76 ± 1.63 μl g⁻¹ min⁻¹ (n = 7; P = 7.2 × 10⁻⁵). There seemed to be larger tear production in P2X₇^−/− animals, and so we re-analysed the data taking sex into account (Fig. 3B and C). Data show that wild-type males and females had similar resting tear secretion, i.e. 2.98 ± 1.47 μl g⁻¹ min⁻¹ and 0.90 ± 0.53 μl g⁻¹ min⁻¹ (n = 9 and 7; P = 0.252). The cholinergic stimulation increased secretion in these animals to 8.79 ± 1.69 μl g⁻¹ min⁻¹ in males and 6.88 ± 2.89 μl g⁻¹ min⁻¹ in females. In P2X₇^−/− animals secretion was significantly higher in males compared to females in the resting state, i.e. 4.41 ± 1.42 μl g⁻¹ min⁻¹ compared to 0.76 ± 0.40 μl g⁻¹ min⁻¹, respectively (P = 0.019); and in the stimulated state it was 15.81 ± 2.07 μl g⁻¹ min⁻¹ for males and 6.21 ± 1.35 μl g⁻¹ min⁻¹ for females (n = 9 and 10; P = 0.001). Notably, as Fig. 3B shows, the P2X₇ receptor knockout in males increased tear secretion by 80% (P = 0.018). For comparison with other glands, the 60 min integrated secretion was 527 ± 237 μl g⁻¹ h⁻¹ and 413 ± 137 μl g⁻¹ h⁻¹ in wild-type males and females; and 924 ± 124 μl g⁻¹ h⁻¹ and 373 ± 81 μl g⁻¹ h⁻¹ in knockout males and females respectively. Note that these apparent high secretory rates are due to the fact that they were corrected for gland weights, i.e. 0.02–0.03 g per two lacrimal glands.

ATP release

The above data show that the P2X₇ receptor seems to be important for fluid secretion in all three types of glands, thus implying that the ligand, ATP, is released within the glands. Our working hypothesis is that cholinergic stimulation of exocrine glands results in release of ATP, which via interaction with P2X₇ receptors leads to co-regulation of exocrine secretion. Firstly, we needed to test that cholinergic stimulation can release ATP in the glands involved, as this has only been shown convincingly for pancreas (Sørensen & Novak, 2001; Kordas et al. 2004; Yegutkin et al. 2006). Secondly, we needed to test that the P2X₇ receptor expression did not have an effect on ATP stores indirectly or directly and in the latter case if it contributes to ATP release as suggested by some studies (Pellegatti et al. 2005). To address these questions, we prepared cell suspensions (containing acini and ducts) from three major exocrine glands of interest – pancreas, parotid gland and lacrimal gland. In one series of experiments, we loaded freshly prepared cells with quinacrine, which is an indicator of ATP stores (Sørensen & Novak, 2001). Figure 4 shows representative fluorescence and transmission images from 6–10 scans made on preparations from all three types of glands obtained from P2X₇^+/+ and P2X₇^−/− animals. Notably, quinacrine was accumulated in granules/vesicles in acini in all gland types. There seems to be no difference between glands obtained from wild-type or receptor knockout animals. That is, the mean fluorescence intensity (grey levels) was 111 ± 5 in acini from wild-type and 108 ± 8 from acini in knockout animals (n = 28 and 7). Another interesting observation was that as in the pancreatic ducts (Sørensen & Novak, 2001), parotid and lacrimal gland ducts showed weaker quinacrine fluorescence in similar regions of interest, and quinacrine was localized in small vesicles as shown in Fig. 4B. That is, the quinacrine intensity in ducts was 46 ± 4 (n = 28), which was significantly lower than in acini (P < 10⁻⁷); and vesicle size was 0.92 ± 0.05 μm in ducts compared to 1.10 ± 0.03 μm in acini (n = 13 and 31; P = 0.003). As shown in the earlier study on pancreatic acini (Sørensen & Novak, 2001), carbachol stimulation resulted in a decrease in the quinacrine fluorescence in luminal regions of acinar cells, indicating that quinacrine/ATP has been released.

In the next series of experiments we set out to determine the ATP release directly using the luciferin/luciferase method. Since we have earlier established that pancreatic cells express ecto-nucleotidases (Yegutkin et al. 2006), nucleotidase inhibitors were included in the cell suspension. First we focused on the effect of cholinergic stimulation on ATP release in wild-type and knockout animals. Figure 5 depicts the results, which show that there was no difference in ATP release in the two groups of animals. This supports the data with the putative ATP-store marker quinacrine (Fig. 4). We also show that other agonists can also elicit ATP release (Fig. 6). Pilocarpine, given at concentrations that we estimated would be the case following i.p. administration in in vivo experiments, caused release of ATP from the three types of glands. Neurotensin had a similar effect in all three glands. CCK-8 also released ATP in pancreas, as established earlier also in in vivo preparations (Yegutkin et al. 2006).

Following ATP release in glands, ATP would activate various P2 receptors, including P2X₇ receptor, if present. Since most G-protein coupled P2Y receptors and cation permeable P2X receptors lead to intracellular Ca²⁺ signals, we measured the change in Fura-2 ratio in cell suspensions from our gland preparations in response to ATP or ATPγS, which is less prone to hydrolysis, and Fig. 7A shows the results. In parotid and lacrimal glands from P2X₇^+/+ and P2X₇^−/− animals, ATP induced similar Ca²⁺ signals. Notably, in pancreatic cells from P2X7^−/− animals, Ca²⁺ signals were significantly lower, i.e. the ΔFura-2 ratio decreased from 0.076 to 0.036 (n = 14 and 16; P = 0.030). This indicates that indeed the contribution from the P2X₇ receptor was lost in pancreas. In addition, we also tested the effect of muscarinic agonist on these gland preparations and the results are shown in Fig. 7B. The simplest assumption is that carbachol stimulation of muscarinic receptors would have led to Ca²⁺ signals due to Ca²⁺ store release and Ca²⁺ influx. Alternatively, or in addition, carbachol could have induced ATP release and subsequent Ca²⁺ signals in the same or other cells in the preparation. The results show that carbachol had similar effects on Ca²⁺ signals in pancreatic cells. However, in both parotid and lacrimal cells, carbachol increased Ca²⁺ signals by 70% and 63% in preparations from P2X₇^−/− animals compared to wild-type animals.

Since there appeared to be some sex differences in exocrine secretion from the three types of glands (Figs 1–3), we examined possible differences in intracellular Ca²⁺ activities. We could not detect any significant difference in pancreatic and parotid cells (data not shown); however, there was a clear difference in lacrimal cells (Fig. 8). Carbachol-induced Ca²⁺ signals were similar in females from P2X₇^+/+ and P2X₇^−/− glands, which were paralleled by similar secretion rates (Fig. 3). However, in cells from P2X₇^−/− male mice, Ca²⁺ signals were about three times higher than in animals with functional P2X₇ receptors (i.e. ΔFura-2 increased from 0.059 ± 0.012 to 0.169 ± 0.023, n = 7 and 8; P = 0.015), and this effect was also reflected in higher tear secretion in these animals (Fig. 3).

ATP release

The fact that pilocarpine stimulation has different effects on secretion in glands from P2X₇^+/+ and P2X₇^−/− animals (Figs 1–3) indicates that P2X₇ receptors are stimulated by ATP released within the glands following cholinergic stimulation. This is confirmed by our experiments on gland cell suspensions, where we show that carbachol and pilocarpine (and other agonists) release ATP from isolated exocrine cells (Figs 5 and 6). Therefore, the theory first proposed and validated for pancreas (Sørensen & Novak, 2001; Novak, 2003; Yegutkin et al. 2006) is also tangible for salivary and tear glands. That is, exocrine cells release/secrete ATP in response to physiological stimuli and this ATP affects the same cells in an autocrine manner or the neighbouring cells in a paracrine manner (e.g. downstream cells such as duct cells in an intact gland).

ATP release mechanisms are currently an issue of intense research and several candidates involved in non-vesicular and vesicular release have been proposed, including connexins, voltage-dependent anion channels, CFTR and even P2X₇ receptor regulated pannexins (Yegutkin, 2008; Praetorius & Leipziger, 2009). Our study shows that ATP release is similar in glands from both wild-type and P2X₇ receptor knockout animals (Fig. 5). This shows that the P2X₇ receptors are not involved in regulation of ATP release alone or in combination with pannexin-1, as proposed for other cells by other workers (Pellegatti et al. 2005; Pelegrin & Surprenant, 2006; Locovei et al. 2007). Quinacrine, an indicator of ATP stores, was distributed in vesicular stores in parotid and lacrimal glands (Fig. 4). This is similar to localization in pancreas described earlier (Sørensen & Novak, 2001), and as we show recently, ATP is accumulated in zymogen granules by vesicular nucleotide transporter VNUT (Haanes & Novak, 2010).

ATP is released into the lumen of glands (Fig. 4B) (Sørensen & Novak, 2001; Sørensen et al. 2003; Ishibashi et al. 2008), as proposed for other epithelia (Schwiebert & Zsembery, 2002; Leipziger, 2003; Novak, 2003), ATP and other nucleotides/sides have been detected in pancreatic juice and also on the eye surface (Sørensen & Novak, 2001; Yegutkin et al. 2006; Crooke et al. 2008). The amount and composition of nucleotides detected on epithelial surfaces or secretions will also depend on ecto-nucleotidase (and kinase) activities, which are high in exocrine gland cells (Sørensen et al. 2003; Yegutkin et al. 2006). Furthermore, the purinergic signalling in the natural microenvironments of the organ may include more components and our simplified experiments do not permit us to conclude whether ATP is also released from nerve endings, or whether ATP release also occurs across the basolateral membranes of epithelial cells.

Secretion – gland by gland

The general patterns of secretion of the three gland types observed in this study are similar to published studies on in vivo and in vitro preparations of glands stimulated by cholinergic or hormonal agonists (Sewell & Young, 1975; Novak & Young, 1986; Walcott et al. 2005; Nakamoto et al. 2009). Regarding the P2X₇ receptor, at first, it may seem surprising that the receptor knockout had a different effect on responses of the three types of exocrine glands. Nevertheless, the glands have different functions, probably a different constellation and regulation of transporters. Purinergic system may be a part of this regulation, and glands can in fact express different P2X₇ receptor splice variants (Nicke et al. 2009).

In pancreas, the major sites of ATP release are acini, which themselves show low functionality of purinergic receptor signalling and apparent lack of P2X₇ receptors (Simeone et al. 1995; Novak et al. 2002). Various P2 receptors, including P2X₇ receptors are expressed and highly functional in rodent and human ducts, where they are associated with Ca²⁺ signals, Na⁺/Ca²⁺ flux and exchange, H⁺/HCO₃⁻ transport and ERK stimulation (Henriksen & Novak, 2003; Hansen et al. 2008; Novak, 2008). In contrast to other exocrine glands, pancreatic ducts are clearly contributing to significant fluid secretion, which is rich in NaHCO₃. Most likely, P2X₇ receptors co-regulate ductal secretion evoked by other agonists, e.g. cholinergic stimulation. In the present study on pancreatic cells (acini and ducts), it is clear that ATP elicited smaller Ca²⁺ signals in cells from P2X₇^−/− animals (Fig. 7), showing that the receptor contributed to this event, and also to secretion that most likely originates in ducts (Fig. 1).

In salivary and lacrimal glands, the main contributors to fluid secretion are acini and they express functional P2 receptors and intracellular signalling (see Introduction). Salivary secretion measurements in the present study were done on whole saliva contributed by the major and minor salivary glands, though the major volume of secretion is contributed by the submandibular and parotid glands and both express P2X₇ receptors (Introduction). As may have been expected then, the P2X₇ receptor knockout reduced salivary secretion (Fig. 2). There are two other relevant secretion studies on salivary glands of similar P2X₇ knockout animals. Nakamoto et al. (2009) studied ex vivo perfused sumandibular gland from wild-type and P2X₇ knockout animals and found that P2X₇ receptor stimulation was responsible for about 70% of fluid secretion, which would support our findings. In contrast, Pochet et al. (2007) measured bulk salivary secretion in wild-type and P2X₇ knockout mice after stimulation with pilocarpine, but could not detect any difference in the first 20 min of fluid secretion between these animals. This anomaly could be explained by our observation showing that the receptor knockout effect in pilocarpine saliva is most apparent in the plateau phase of secretion (after some 20 min). Interestingly, in this phase of secretory response, the saliva secretion is driven by ion transporters associated with H⁺/HCO₃⁻ transport rather than the Na⁺–K⁺–Cl⁻ and Cl⁻ transporters (Novak & Young, 1986). Thus, we can predict some parallels between pancreatic and salivary gland secretions – in regulation and cellular mechanisms.

The aqueous layer of the tear film contains proteins, electrolytes and water, which are mainly secreted by the lacrimal gland; minor secretion containing lipids is mainly by the Meibomian glands; and mucus is produced by epithelial and goblet cells (Crooke et al. 2008). It is reasonable to assume that tear fluid secretion would be mainly contributed by lacrimal gland cells, which we used in this study. The cellular model for secretion is well studied on isolated acini (and the latest also with ducts) and is similar to salivary and pancreatic acini (Dartt, 2009). A recent paper by Dartt's group (Hodges et al. 2009) shows that P2X₇ receptors are expressed in lacrimal acini of male rats – where they stimulate Ca²⁺ signals, ERK1/2 and protein secretion. It is perhaps unexpected then that our study shows that P2X₇ receptor knockout in mice had no effect on fluid secretion (females) or even increased fluid secretion (males) (Fig. 3). The simplest proposal is that protein and fluid secretions are not regulated in parallel by the P2X₇ receptors. Another suggestion is that there are interactions between cholinergic (muscarinic receptor) and purinergic signalling. This interaction may be partly reflected by Ca²⁺ signals in all glands. Although the Ca²⁺ signal measurements are relatively crude in our study, as we monitored global peak changes in cell suspensions containing acini and ducts, we already see some interesting patterns worth discussing below.

Interaction of purinergic and cholinergic system and calcium signals

Interaction between purinergic receptors and other receptors can occur at several levels: at the plasma membrane level, via interaction of intracellular signalling pathways, and at the organ level. On the level of plasma membrane, G-protein coupled receptors can heteromerize and also ligand-gated receptors, such as P2X and nicotinic receptors, exhibit physical cross-talk (Khakh et al. 2000; Koles et al. 2008). At the organ level, Tenneti et al. (1998) have shown that there is trans-synaptic regulation of receptor-mediated signalling between muscarinic and P2X receptors, such that removal of the autonomic innervations increases ATP responses (presumably via P2X₄ and P2X₇ receptors) in parotid acinar cells. From our study it seems, though, that this is not the other way around, i.e. removal of P2X₇ receptors did not lead to better secretion by cholinergic stimulation (at least in the salivary glands and pancreas), although Ca²⁺ signals were increased (in salivary glands).

In pancreas, clearly knockout of the P2X₇ receptor led to lower Ca²⁺ signals (in response to ATP) and secretion was decreased. This secretion most likely originated in ducts rather than acini, as the former contribute significantly to secretion and express many functional P2 receptors, while acini do not (Hede et al. 1999; Novak et al. 2002). In parotid and lacrimal cells, P2X₇ receptor knockout had no significant effect on ATP-mediated Ca²⁺ signals in cell preparations, most likely because these acini express many functional P2 receptors that can substitute for P2X₇ receptor knockout (see Introduction). Nevertheless, the receptor knockout led to increased Ca²⁺ signals in response to muscarinic receptor stimulation. This is in accordance with findings of Hurley et al. (1993) on submandibular gland acini, which showed that P2X₇ receptors prevented Ca²⁺ release from intracellular stores by carbachol and also adrenaline. Similar findings were made by Jørgensen et al. (1995) on parotid acini, where the P2Z/P2X₇ receptor diminished the effect of the muscarinic receptor on release of inositol 1,4,5-trisphosphate and store Ca²⁺. Therefore, it was expected that P2X₇ knockout would improve Ca²⁺ signals (Fig. 7). But receptor knockout did not rescue pilocarpine-induced total salivary gland secretion in intact animals; in fact it decreased it (Fig. 2). Though for the isolated submandibular glands from mice of both sexes, Nakamoto et al. (2009) reported that P2X₇ receptor knockout increased muscarinic receptor induced fluid secretion. This latter observation is more in accordance with what we propose for the lacrimal glands. That is, in lacrimal glands, P2X₇ receptor knockout clearly improved both Ca²⁺ signalling and secretion, particularly in the males (Figs 3 and 8).

Taken together, in some glands the muscarinic and P2X₇ receptors interact, most likely via signalling pathway that converge and decrease Ca²⁺ signals, which eventually increases secretion in salivary glands but decreases it in lacrimal glands. It is then important to consider that increased Ca²⁺ signals may not be synonymous with increased secretion. This anomaly is also seen with P2Y₂ receptor stimulation, which leads to large Ca²⁺ signals, yet inhibits specific K⁺ channels (BK) and secretion (Hede et al. 1999). Thus, it will be necessary to look beyond global Ca²⁺ signals (detected in this study) to specific exocrine cells (acini and ducts), specific ion transporters and localized subcellular signals and involvement of other signalling macromolecules, such as mitogen activated protein kinases and members of the PKC family that can up- and down-regulate plasma membrane transporters (Rosse et al. 2010).

Sex differences

In this study we detected sex differences in exocrine secretion (Figs 1 and 3). Sex hormones affect development, structure and function of many organs, including epithelia (Sabolic et al. 2007). Regarding morphology, there is well documented sexual dimorphism in rodent salivary glands, where male glands contain well developed granular convoluted tubules containing granules with growth factors and kallikrein (Barka, 1980; Jayasinghe et al. 1990; Tandler et al. 2001). Male glands also contain smaller volume of acini, which may explain their lower secretion (Fig. 2). Regarding sexual dimorphism and function, some of the hormonal effects include rapid non-genomic responses by influencing intracellular Ca²⁺, PLC and various kinases (PKC, PKA, MAPK) and ion channels. Relevant to our study may be the finding that 17β-oestradiol inhibits Cl⁻ secretion in colonic crypts, and this seems to be exerted via inhibition of a K⁺ channel, KCNQ1, that is regulated via PKCδ and PKA activation (O’Mahony et al. 2007). Similar mechanism could explain lower secretion in female pancreas and tear glands compared to male glands.

Most interesting is the observation that we also detect difference in the effect of P2X₇ receptor knockout in male and female glands. Most importantly, the P2X₇ receptor knockout had no effect on secretion in all female exocrine glands. In males, though, the receptor knockout decreased secretion in pancreas (and salivary glands) but increased it in lacrimal glands. These effects do not seem to be related to ATP release (Fig. 5), but may be directly related to the P2X₇ receptor, such that in females, for example, oestradiol would inhibit the P2X₇ receptor. There are a few studies that support our observations regarding the issue of P2X₇ receptor and sex differences and/or oestradiol. In studies on bone remodelling in mice, the P2X₇ receptor knockout had less obvious effects on the loss of bone mass of females compared to males (Ke et al. 2003). In post-menopausal women oestrogen treatment prevented loss of bone mass (Ohlendorff et al. 2007). On the cellular level, two studies indicate that the P2X₇ receptor function is partly regulated by oestrogens. In studies on monkey kidney COS cells expressing the hP2X₇ receptor, but not the oestrogen receptor, it was shown that 17β-oestradiol had rapid inhibitory effects on cation flux through the receptor (Cario-Toumaniantz et al. 1998). A similar inhibitory effect of oestrogen on Ca²⁺ influx, and also on apoptosis associated with the pore formation, was observed in uterine cervical cells (Gorodeski, 2004).

In conclusion, our study shows that the P2X₇ receptors have effects on secretion of three major exocrine glands and we propose that they play a role in physiological regulation. We show that the signalling molecule, ATP, can be released from glands following cholinergic stimulation and this is independent of P2X₇ receptors. Furthermore, we show that P2X₇ receptors act in conjunction with muscarinic receptors to increase secretion (pancreas, salivary glands) and to decrease secretion (tear glands) and these latter effects are related to male/female origin of the gland.